Jeffrey Sullivan, MS

 We examined how a population susceptible to hepatitis C virus (HCV) moves through the HCV screening and linkage-to-care (SLTC) continuum across insurance providers (Medicare, Medicaid, commercial) and identified opportunities for increasing the number of patients who complete the SLTC process and receive treatment.

 A cohort of 10,000 HCV-susceptible patients was simulated through the HCV SLTC process using a Markov model with parameters from published literature. Three scenarios were explored: baseline, in which each step required a separate visit and all infected saw a specialist; reflex, which reflexed antibody and RNA testing; and consolidated, which reflexed antibody, RNA, fibrosis staging, and genotype testing into 1 step, with an optional specialist visit. For each scenario, we estimated the number of patients lost at each stage, yield, and cost.

 Streamlining the SLTC process by reducing the number of required visits results in more patients completing the process and receiving treatment. Among antibody-positive patients, 76% of those with Medicaid and 71% of those with Medicare and commercial insurance are lost to follow-up in baseline. In reflex and consolidated, these proportions fall to 26% and 27% and 4% and 5%, respectively. The cost to identify and link 1 additional infected patient to care ranges from $1586 to $2546 in baseline and $212 to $548 in consolidated. Total cost, inclusive of treatment, ranges from $1.0 million to $3.1 million in baseline and increases to $3.8 million to $15.1 million in reflex and $5.3 million to $21.0 million in consolidated.

 Reducing steps in the HCV SLTC process increases the number of patients who learn their HCV status, receive appropriate care, and initiate treatment.

This study evaluated the impact of streamlining the hepatitis C virus (HCV) screening and linkage-to-care (SLTC) process on costs, yield, and patients lost to follow-up by integrating reflex testing for early steps in the process. The findings are relevant for clinicians and managed care decision makers involved in HCV SLTC programs.

Reducing the number of required visits during the SLTC process decreases the number of patients lost to follow-up by 62% to 95%.

Streamlining the HCV SLTC process results in more patients who are aware of their HCV status, receive appropriate care, and are ultimately treated.

Up to 3.5 million people in the United States are infected with chronic hepatitis C virus (HCV), and half are unaware of their infection status.

 Current guidelines recommend 1-time HCV screening for individuals born between 1945 and 1965 and individuals with increased risk of infection, but initial screening represents only the first stage in the screening and linkage-to-care (SLTC) process.

 To detect chronic infection, patients with a positive HCV antibody test must have confirmatory RNA testing, and for those with chronic HCV, additional diagnostics, including genotype testing and fibrosis staging, are recommended before treatment.

Reflex testing, in which RNA is tested immediately following a reactive antibody test using the same blood draw, represents a simplified SLTC process and allows patients to definitively know their HCV status following 1 visit.

 Without reflex testing, an estimated 33% to 47% of patients who receive a positive antibody test do not receive confirmatory RNA testing, highlighting the importance of a streamlined process for patient awareness.

Although fewer visits in the SLTC process may result in fewer patients lost to follow-up, other barriers may result in patients dropping out of the process prior to initiating treatment. For example, HCV guidelines still recommend subspecialty consultation for patients with advanced fibrosis or cirrhosis.

 The need for specialty care may disproportionately impact patients with less access to care, such as those who use community health centers.

 Even if patients successfully complete all screening and diagnostic testing and receive a prescription for treatment, they still may not be treated if their payer policy includes coverage restrictions, such as prior authorization (PA).

Given that only 16% of chronically infected patients are eventually prescribed treatment,

 it is important to identify steps in the SLTC process where patient retention is lowest and improve retention at those points. To examine this issue, we developed a model that simulates the HCV SLTC process from antibody testing through treatment initiation. The minimum number of visits required prior to a treatment decision varied from 2 to 4, and the resulting costs, yield, and patients lost to follow-up were estimated depending on patients’ insurance provider (Medicaid, Medicare, or commercial).

A discrete-time Markov model was developed to simulate the HCV SLTC process and was stratified by insurance type: Medicaid, Medicare, and commercial. The model follows 10,000 patients from antibody screening through treatment initiation until 1 of 5 conditions is met: (1) they are found not to have chronic HCV, (2) a “no treatment recommended” decision is made, (3) PA is denied, (4) treatment is initiated, or (5) they drop out before meeting any of the prior conditions and are lost to follow-up (henceforth, “lost”).

 Patients enter the model and receive an antibody test. Those who are antibody-negative are not infected with HCV, require no additional testing, and have completed the SLTC process. Patients who are antibody-positive (Ab+) continue to confirmatory RNA testing or are lost.

RNA testing assesses the presence of chronic infection. Patients who are RNA-negative have no active infection and have completed the screening process. Patients who test RNA-positive and have been infected longer than 6 months are chronically infected and continue to a specialist for further testing or are lost.

Patients with chronic HCV receive genotype testing and noninvasive liver fibrosis staging at a specialist visit; results determine their treatment regimen and duration. Biopsies account for fewer than 10% of fibrosis staging tests

 and were excluded from our model. At this stage, patients either receive a “no treatment recommended” decision, are prescribed treatment, or are lost. Patients who receive a decision of no treatment recommended have completed the screening process.

If treatment is recommended, patients transition through additional stages before receiving therapy. We do not explicitly model additional tests that may follow the treatment recommendation, such as NS5A resistance testing or renal function testing; however, such tests could be conducted during the specialist visit, which avoids additional visits during the SLTC process. At least 29 states require some duration of sobriety for Medicaid patients.

 Therefore, Medicaid enrollees in our model must meet sobriety requirements and obtain PA before initiating treatment. Commercial and Medicare enrollees must obtain PA.

Model parameters were drawn from the published literature and stratified by insurance type where possible. Key parameters for the model include transition probabilities, average timing values, and per visit costs. HCV prevalence plays a key role in model outcomes because it determines the number of Ab+ patients who progress beyond the initial state. Of our 3 insurance strata, Medicaid has the highest prevalence of Ab+ patients (16%), followed by Medicare (9%) and commercial (5%).

 The proportion of Ab+ patients with chronic HCV is the same for all strata (79.7%).

) provides a full description of the model assumptions and parameters.

The baseline scenario previously described requires at least 4 visits before chronically infected patients receive a treatment recommendation (Figure 1). We also estimated reflex and consolidated scenarios in which chronically infected patients require a minimum of 3 or 2 visits, respectively, before treatment recommendation. All stages of the SLTC process in baseline are included in reflex and consolidated, but they are condensed to varying degrees. Beyond the treatment decision step, the 3 scenarios are identical.

Baseline includes the SLTC steps recommended in current HCV guidelines, with each step in the process requiring a separate visit and only specialists providing genotype testing, fibrosis staging, and treatment decisions. Baseline thus requires 4 visits for a chronically infected patient to receive a treatment recommendation and is the least efficient scenario.

Reflex also includes the SLTC steps recommended in current HCV guidelines, but reflexes antibody and RNA testing so that 2 blood samples are collected from a single draw at the first visit and if the first is Ab+, the second is automatically tested for HCV RNA.

 Thus, the process can be completed in 3 visits rather than 4.

Reflex also eliminates the specialist visit for less clinically complex patients who can be effectively managed by primary care physicians (PCPs). To be conservative, we assumed that only patients with fibrosis scores below F2 can be managed by PCPs, which resulted in 60% of patients requiring a specialist visit.

Consolidated represents a hypothetical best-case scenario in which all tests are reflexed and a specialist visit is not required for patients with fibrosis scores below F2. This scenario requires at least 2 visits for chronically infected patients to receive a treatment decision and provides the fewest opportunities for patients to be lost.

Although pangenotypic treatment is now available, guidelines still recommend genotype testing, and all 3 scenarios include it. Genotype testing can also be used to guide treatment for patients who do not receive pangenotypic treatment. To reflex noninvasive fibrosis staging, diagnostic tests using a blood draw (eg, AST [aspartate aminotransferase] to Platelet Ratio Index or FibroTest) would be required.

 Although these may be uncommon as the primary means for fibrosis staging in current practice, they are feasible. Current HCV treatment guidelines recommend combined blood- and image-based fibrosis testing, so our consolidated scenario should be considered exploratory, as an estimate of potential benefits from an SLTC process that minimizes patient visits.

We estimated several outcomes for each scenario and insurance type: number of patients lost at each stage, yield (percentage of patients entering the model who complete the process and initiate treatment), conditional yield (percentage of patients with chronic HCV who complete the process and initiate treatment), and several cost outcomes. Total screening costs include the cost of antibody testing, RNA testing, genotype testing, fibrosis staging, and, when relevant, specialist and sobriety costs. Total costs include screening costs plus the cost of treatment. We assumed that treatment cost equals the wholesale acquisition cost of sofosbuvir 400 mg/velpatasvir 100 mg (Epclusa) discounted by 46%.

 To estimate the cost to identify and link 1 additional patient to care, we calculated the number of patients screened to yield 1 patient in the genotype/fibrosis step of the model and calculated the cost of antibody and RNA testing for those patients. This outcome provides an additional measure of efficiency of the SLTC process prior to receiving a treatment recommendation and is independent of the number of patients treated. Other outcomes are presented in the eAppendix, including cost per person screened, timing-related results, and yield conditional on the number of Ab+ patients.

SLTC results for each strata and scenario are presented in 

. The figures show the number of patients lost after each step; fewer patients lost indicates a more efficient process. In baseline, 12% of Medicaid, 6% of Medicare, and 4% of commercial patients are lost before treatment. Across all insurance types, most patients are lost after RNA testing. For reflex and consolidated, 4.0% and 0.7%, respectively, are lost before treatment for Medicaid, 2.0% and 0.5% for Medicare, and 1.0% and 0.3% for commercial patients.

Yield and total lost results are presented in the 

. Yield incorporates both HCV prevalence in the screened population and the likelihood of loss. High yields therefore result from high prevalence, screening process efficiency, or both. Baseline yields are 0.5%, 0.7%, and 0.2% for Medicaid, Medicare, and commercial patients, respectively. The higher efficiency of the reflex and consolidated models translates into higher yields; reflex yields are 3.5%, 3.1%, and 0.9%, respectively, and 4.9%, 4.4%, and 1.2% in consolidated.

Conditional yield describes the efficiency of the screening process for chronically infected patients. Baseline conditional yields are 4%, 9%, and 5% for Medicaid, Medicare, and commercial, respectively, increasing to 28%, 44%, and 22% in reflex and 31%, 49%, and 24% in consolidated.

The Table also presents total costs, total screening costs, screening costs per patient treated, and the cost to identify 1 additional patient and link them to care. We used undiscounted costs in our model; therefore, our cost results represent an upper bound.

Total costs are driven by the total treated, and in baseline are highest for Medicare ($3.1 million) and lowest for commercial ($1.0 million); total costs increase substantially for reflex and consolidated, ranging from $3.8 million to $15.1 million and $5.3 million to $21.0 million, respectively. Although commercial has the lowest total screening cost, it also treats the fewest patients, leading to higher per person costs. In baseline, the cost per person treated is $7843 for Medicaid, $4833 for Medicare, and $14,176 for commercial.

For baseline, the cost to identify 1 additional chronically infected patient and link them to care is highest in the commercial population ($2546) and lowest in the Medicare population ($1539), but Medicaid sees the largest reductions in that cost when the process is collapsed from baseline to reflex or consolidated.

Three alternative analyses were conducted: (1) “fixed prevalence,” which assumed the same HCV Ab+ prevalence across insurance types; (2) “no genotype,” which removed genotype testing; and (3) “no sobriety requirements,” which removed sobriety requirements. Results for alternative analyses are presented in the eAppendix.

“Fixed prevalence” allows us to compare the efficiency of the SLTC process across insurance types while holding constant population disease prevalence. Commercial has the highest per person costs in the main analysis, but Medicaid and commercial costs are similar in the fixed prevalence analysis, suggesting that lower per person costs in Medicaid are driven primarily by higher prevalence.

Although genotype testing is still recommended in HCV treatment guidelines, “no genotype” explores the impact of removing genotype testing, which may be possible with pangenotypic therapies. Because genotype testing occurs during the same visit as fibrosis staging, removing it reduces per person costs by $351 but does not reduce visits.

“No sobriety requirements” removes a barrier to treatment initiation that occurs late in the SLTC process and affects only the Medicaid population. Removing sobriety testing reduces screening costs by $48 per drug test.

Comparing baseline, reflex, and consolidated results shows the value of streamlining the SLTC process. Collapsing baseline to reflex reduces the number of patients lost prior to receiving a treatment recommendation by 62% to 66%. Further streamlining to consolidated reduces the number lost by 92% to 95%. Because the Medicaid population has the most inefficient SLTC process to begin with, it experiences the largest improvements from streamlining the process.

Although our analysis focuses on the number of visits as a measure of SLTC process efficiency, the underlying prevalence in a given population also plays an important role in the process. For example, although the Medicaid subpopulation loses the most patients, it also has the highest Ab+ prevalence (16%), resulting in higher yield versus the commercial population. Moreover, in consolidated, only 20 additional Medicaid patients need to be screened on average to get 1 additional chronically infected patient into treatment compared with 100 additional patients screened to achieve the same result in the commercial population.

Our findings are consistent with those of a recently published study of the care continuum for patients with HCV diagnosed in 2 urban emergency departments.

 In the study, the Medicaid process was less efficient, with only 8.5% of RNA-positive patients initiating treatment compared with 12% of Medicare patients. Additionally, 82% of Ab+ patients completed viral load testing, similar to our finding of approximately 80% of patients progressing to RNA testing. Generally, the literature presents conditional yield rather than yield; conditional yield estimates range from 3.9% to 24%.

 Our baseline conditional yields (3.7%-9.5%) are on the lower end compared with recent studies. Our baseline Ab+ conditional yields (eAppendix) range from 2.7% to 7.5%, which is consistent with findings from 2 recent studies (3.3% and 4.0%).

Not surprisingly, total costs increase dramatically from baseline to reflex and consolidated, because more patients receive treatment. This paper is not intended as a cost-benefit exercise, nor do we model other medical expenditures for patients with HCV. However, the increased treatment costs are arguably of high value because identifying and treating more patients will provide benefits associated with reduced transmission rates, long-term cost savings on medical expenditures related to untreated HCV, and a reduction in liver transplants.

Reducing the number of patients lost decreases screening costs per person treated because the total system costs are spread among more patients. This aligns with prior literature showing that expanded HCV screening provides the most value when coupled with expanded treatment.

 Additionally, there are costs associated with the SLTC process that are difficult to measure (eg, patient navigation, social work) and are not included in our analysis. It is likely that Medicaid patients would benefit most from these services and incur additional costs, but this population also experiences the greatest gains from reflex and consolidated.

For all insurance types, a majority of patients are lost prior to visiting a specialist, which suggests that having insurance does not eliminate inefficiencies associated with multiple visits required in the SLTC process. Although our 3 scenarios focus on streamlining the SLTC process prior to treatment recommendation, barriers to treatment exist in later stages of the process. Specifically, PA poses a significant barrier for patients who are prescribed treatment, particularly in the Medicaid and commercial populations. Of patients who seek PA, 46% and 55% of Medicaid and commercial patients, respectively, are denied, whereas only 13% of Medicare patients are denied.

A patient who is denied PA is comparable with one who is lost, and for patients who are eventually denied, streamlining the process simply delays the point at which they are lost. This delay increases overall costs from screening and time spent in the process but does not change the disease outcome because treatment is not received. Consequently, to maximize the number of patients treated, barriers to treatment must be reduced.

Articles

An optimized hepatitis C virus screening and linkage-to-care process reduces the number of patients lost to follow-up and improves linkage to care for Medicare, Medicaid, and commercially insured patients.

Evaluating HCV Screening, Linkage to Care, and Treatment Across Insurers

 To compare the well-being of long-term cancer survivors with that of US residents of similar age and demographic characteristics, patients recently diagnosed with cancer, and individuals with chronic illness.

 Using the Health and Retirement Study, a survey of US residents older than 50 years, we defined 4 cohorts: long-term cancer survivors (>4 years post diagnosis), individuals recently diagnosed with cancer (≤4 years post diagnosis), individuals with chronic illness, and US residents older than 50 years (“nationally representative cohort”). Well-being measures included self-reported health, utility, happiness, medical utilization and spending, employment, and earnings, and these measures were compared across cohorts, adjusting for survey year, demographic characteristics, smoking, and number of comorbidities. We imputed medical spending using the Medical Expenditure Panel Survey and the Medicare Current Beneficiary Survey. 

 Long-term cancer survivors fared significantly better than those recently diagnosed with cancer, those with chronic illness, and individuals in the nationally representative cohort in the majority of well-being measures (

<.05), including fewer doctor visits, hospitalizations, and hospital nights; better utility and self-reported health; and greater likelihood of employment. Long-term cancer survivors had lower healthcare spending than those recently diagnosed with cancer (

<.01) and significantly greater happiness than the nationally representative cohort and those with chronic illness (

 Although patients with cancer experience diminished well-being in the short term across a variety of measures, in the long term, cancer survivors do as well as or better than US residents of similar age and demographic characteristics. This finding is striking given that one might expect long-term cancer survivors to do worse than similar individuals without a history of cancer.

Although patients with cancer experience diminished well-being in the short term across a variety of measures, in the long term, cancer survivors do as well as or better than US residents of similar age and demographic characteristics.

We found that long-term cancer survival has an effect on well-being comparable with that of many chronic conditions. This is notable given that 70% of US residents older than 50 years have at least 1 comorbidity.

Advances in cancer care that offer a chance of long-term survival may enable patients to experience a high quality of life.

Cancer care is changing rapidly. Over the past 60 years in the United States, population aging and declining mortality from heart disease have led to an increase in the total number of deaths from cancer.

 However, this increase in the total number of deaths masks substantial improvement in the cancer mortality risk faced by individuals. Adjusting for population aging, cancer mortality rates in the United States have declined by 23% from 1990 to 2012, from 214.95 to 166.40 per 100,000 individuals.

 Reasons for this decline include advances in early detection and treatment and decreased smoking.

 Although the debate continues about the efficacy of the “war on cancer” and further progress is needed,

 the trend of improving mortality is clear. Studies examining many tumor types have found that advances in care have led to significant mortality declines in recent decades.

 These reductions in mortality have increased the number of cancer survivors.

However, surviving cancer does not guarantee a high quality of life. Adverse effects of cancer treatment are well documented. Cardiac toxicities are a known complication of specific chemotherapeutics used in breast and other solid tumors, leaving many cancer survivors with degrees of heart failure.

Radiotherapy in patients with breast cancer has been linked to an increased risk of coronary events.

 Impairment of renal function is a complication of certain cancer treatment regimens, often resulting in acute damage and chronic kidney disease.

Patients with cancer also face an increased risk of second primary malignancies.

 For example, a meta-analysis found that patients treated for Hodgkin lymphoma are at increased risk of developing second primary lung cancer.

 A study of testicular cancer survivors found that, compared with surgery only, radiotherapy below the diaphragm significantly increased the risk of second primary cancers or cardiovascular diseases.

Because long-term cancer survivors face the risk of second primary malignancies and may live with complications from treatment toxicities, one might expect that they would experience a low quality of life. However, research to date has tended to focus on clinical and short-term outcomes, while comparatively little is known about the well-being of long-term cancer survivors. As improvements in treatment have made long-term survival a possibility for some patients with advanced cancers, understanding the well-being of long-term survivors is of increasing importance. 

In this study, we measured the well-being of long-term cancer survivors and compared their well-being with that of individuals recently diagnosed with cancer, individuals with chronic illness, and US residents of similar age and demographic characteristics. According to the CDC, “There is no consensus around a single definition of well-being.”

 Therefore, we studied a variety of well-being measures encompassing patients’ health, life satisfaction, and productivity.

This study compared well-being outcomes across 4 cohorts: 1) long-term cancer survivors, 2) patients recently diagnosed with cancer, 3) individuals with at least 1 nonterminal chronic illness, and 4) a nationally representative sample of US residents older than 50 years. Regression analysis was used to compare outcomes across the cohorts, adjusting for survey year, demographic characteristics, smoking, and number of comorbidities. A study schematic is presented in 

We used the Health and Retirement Study (HRS), a nationally representative biannual survey of US residents older than 50 years, to construct the 4 cohorts and follow their well-being outcomes over time. The HRS was designed to illuminate various aspects of US residents’ lives, including health, finances, and employment, as they approach and move into retirement.

The common definition of long-term survival in individuals who have had a cancer diagnosis is 5 years post diagnosis.

 Because the HRS data are only collected every 2 years, we chose 4 years as the threshold for separating short- from long-term cancer survivors (ie, long-term survivors were diagnosed at least 4 years ago, whereas short-term survivors were diagnosed less than 4 years ago). This definition enabled us to focus on long-term survivors beyond the acute treatment phase. Taking a long-term perspective was important because research suggests that the transition from patient to long-term survivor can be challenging,

 and we sought to understand survivors’ well-being after that transition has been undertaken.

The chronic illness cohort included individuals with a diagnosis of diabetes, hypertension, heart disease, lung disease, or prior stroke, but with no prior cancer diagnosis. The nationally representative cohort consisted of all individuals older than 50 years in the 2010 HRS data, as this was the most recent nationally representative wave in the HRS. The cohorts were analyzed over 2004 through 2012 at the person-year level, although we used HRS data from as far back as 1992 to identify whether patients were short-term or long-term cancer survivors. A patient could appear in multiple cohorts. 

Because most outcomes in the HRS data are self-reported, these data are not ideally suited for tracking medical expenditures. Therefore, we developed an algorithm to impute total medical expenditures, including medical and pharmacy costs, in the HRS based on 2 data sets: the Medicare Current Beneficiary Survey (MCBS) for those older than 65 years and the Medical Expenditure Panel Survey (MEPS) for those between ages 50 and 64 years.

Specifically, we used the MCBS Cost and Use files, which combine results of a nationally representative survey of Medicare beneficiaries with administrative data to obtain accurate estimates of Medicare program spending.

 Health conditions were identified using the Health Status and Functioning (Community) data file from the Cost and Use files. We used the MCBS and the HRS from 1992 to 2010 to develop this imputation and applied it to our 2004-2012 HRS sample.

To estimate the medical expenditures of individuals aged between 50 and 64 years, we used the MEPS years 2000 through 2010. Although MEPS is based on survey data, it is designed to measure healthcare cost and utilization.

 In addition, the MEPS data were used to estimate utility for everyone older than 50 years based on the EQ-5D-3L measure, a utility measure introduced by the EuroQol Group in 1990. All health conditions were identified via self-reports from the MEPS condition files and household consolidated files.

Because there is no single consensus definition of well-being,

 we selected a wide range of well-being outcomes, including medical service use, healthcare spending, employment status, earnings, self-reported health, utility, and happiness.

Medical service use measures included the numbers of doctor visits, hospital admissions, and hospital nights over the 2 years prior to the survey. Healthcare spending (medical and pharmacy) was measured annually in the year of the survey and inflated to 2014 dollars using the medical Consumer Price Index (CPI). We constructed a binary variable for whether an individual was working at the time of the survey. We identified individuals as working if they were working full-time, they were working part-time, or they were workingpart-time in retirement. Earnings were measured annually in the year of the survey, conditional on the individual working, and inflated using the CPI for all urban consumers. Earnings summarized the market value of an individual’s effort in the labor force and served as a proxy for productivity. 

For self-reported health, we constructed a binary variable: poor or fair health versus good, very good, or excellent health. Self-reported health was measured at the time of the survey.

Utility was measured using a regression of EQ-5D-3L score on health, demographics, and functional status over the year prior to the survey.

 For happiness, we constructed a binary variable: not at all, a little, or moderately happy versus quite a bit or very much happy. Happiness was measured over the 30 days prior to the survey. 

We looked at the differences in these well-being outcomes across the 4 cohorts controlling for the year of the survey (2004-2012), age group (51-54, 55-59, 60-64, 65-69, 70-74, 75-79, 80-84, ≥85 years), gender, race (white/Caucasian, black/African American, other), ethnicity (Hispanic, non-Hispanic), education (less than high school, general educational development certificate, high school graduate, some college, college and above), health insurance (insured, not insured), smoking history (ever smoked, never smoked), and comorbidity count. The latter included any prior diagnoses of cancer, diabetes, heart disease, hypertension, lung disease, or stroke. Prior diagnosis was determined by whether an individual indicated ever being diagnosed with any of the mentioned diseases, as determined in the year of eligibility for the cohort in question. 

We performed a descriptive analysis on the HRS data using standard weights to compare outcomes across cohorts at the person-year level. Simple 2-way 

tests were conducted to assess the significance of the differences in means across cohorts. 

We ran multivariable analyses for all outcomes to compare the well-being of long-term cancer survivors with that of the other 3 cohorts, adjusting for survey year, demographic characteristics, smoking, and number of comorbidities. Because cancer could plausibly affect the subsequent development of comorbidities,

 we also ran an alternative specification omitting number of comorbidities as a covariate. Models were selected to be appropriate for the given outcome. We used a Poisson model for medical service use, an ordinary least squares model for annual earnings conditional on being employed, a Tobit model for utility, and a generalized least squares model with a gamma distribution and log link for healthcare spending. Additionally, we used logistic models for the binary outcomes: self-reported health, happiness, and employment status. After conducting the analyses, we estimated the predicted values of all outcomes using the mean values of the covariates for the nationally representative cohort in 2010. 

In 2010, there were 1184 long-term cancer survivors; 676 individuals recently diagnosed with cancer; 12,583 individuals with chronic illness; and 22,034 US residents older than 50 years (“the nationally representative cohort”). Over the biennial survey waves from 2004 to 2012, there were 8817; 3374; 57,108; and 22,034 person-years in the cohorts, respectively. Cohort characteristics are shown in 

. Long-term cancer survivors were, on average, older compared with the other cohorts and more likely to be female, white, and non-Hispanic and to have health insurance. 

presents the results of the descriptive analysis. The means of medical service use, employment, and medical spending for individuals recently diagnosed with cancer, the chronic illness cohort, and the nationally representative cohort were all significantly different from those of long-term cancer survivors (

<.01). The means of utility and earnings conditional on being employed for the chronic illness and nationally representative cohorts were significantly different from those of long-term cancer survivors (

<.01), whereas those of the recently diagnosed cohort were not statistically different from those of long-term cancer survivors. Grouping self-reported health as excellent, very good, or good versus fair or poor, the recently diagnosed cohort had a significantly lowerlikelihood of being excellent/very good/good,and the nationally representative cohort had a significantly higher likelihood, than long-term cancer survivors (

<.01). Grouping happiness as very much or quite a bit versus moderately, a little, or not at all, the chronic illness cohort was significantly less likely to be very much/quite a bit than long-term cancer survivors (

The results of the multivariable analyses are presented in 

, which shows the predicted values of the different well-being measures in each cohort. Long-term cancer survivors fared better than the recently diagnosed cohort in terms of healthcare utilization, utility, healthcare spending, self-reported health, and employment (all 

<.01). Differences in other outcomes were not significant. Similarly, long-term cancer survivors fared better than individuals with chronic illness in terms of healthcare utilization, utility, self-reported health (all 

<.01), happiness, and employment status (both 

<.05). Differences in other outcomes were not significant. Compared with the nationally representative sample, long-term cancer survivors fared better in terms of healthcare utilization, utility, employment status (all 

<.01), self-reported health, and happiness (both 

<.05). Other outcomes were not significantly different. 

The multivariable results are presented graphically in 

, which shows how the well-being measures compared across the 4 cohorts. To display all outcomes on a common scale, well-being measures were normalized so that the value of the long-term cancer survivors cohort is 100%.

The results of the alternative specification in which number of comorbidities was omitted as a covariate are presented in the 

). As in the base case, long-term cancer survivors fared better than the recently diagnosed cohort in the majority of well-being outcomes, with the exceptions of utility, happiness, and earnings, which were not significantly different. Compared with individuals with chronic illness, long-term cancer survivors had moderately higher healthcare utilization and spending and lower self-reported health but greater utility. There were no significant differences in hospital nights, happiness, employment, and earnings between long-term cancer survivors and individuals with chronic illness. Compared with the nationally representative cohort, long-term cancer survivors had greater healthcare utilization and spending and lower self-reported health and employment. Differences between long-term cancer survivors and the nationally representative cohort in terms of utility, happiness, and earnings were not significant.

Our study results show a marked improvement in cancer survivors’ well-being in the long term compared with the first 4 years after diagnosis. Moreover, long-term cancer survivors fared at least as well as average US residents across a variety of well-being measures, controlling for demographics and the number of comorbidities. When comorbidities were not controlled for, long-term cancer survivors had modestly higher healthcare utilization and spending and modestly lower self-reported health and employment compared with average US residents. Taken together, these results suggest that when long-term cancer survivors fare worse than individuals without cancer, it is due to their comorbidity burden. These differences tended to be modest in size (eg, 1 extra doctor visit over 2 years). Moreover, even without controlling for comorbidity burden, there were no differences in happiness, utility, or earnings between long-term cancer survivors and average US residents.

Our analysis indicated that 70% of US residents older than 50 years have at least 1 comorbidity. Therefore, the fact that long-term cancer survivors fare comparably with others their age, many of whom are also dealing with health conditions, suggests that in the long term, cancer has an effect on well-being comparable with that of many chronic conditions.

Previous literature has compared the well-being of cancer survivors with that of individuals with chronic illness.

 Steel et al studied the difference in health-related quality of life (HRQoL) of individuals with hepatocellular carcinoma (HCC), individuals with chronic liver disease (CLD), and the general population.

 Those with HCC had statistically significantly poorer HRQoL than those with CLD and the general population. Those with HCC and CLD had better social and family well-being than the general population. 

In a similar study, Elliot et al compared the well-being of cancer survivors without other chronic conditions, cancer survivors with chronic conditions, and individuals with chronic conditions other than cancer with that of healthy individuals.

 They found that, generally, the well-being outcomes of cancer survivors without chronic conditions did not statistically differ from those of individuals with 1 chronic condition. 

Other studies have compared the well-being of cancer survivors with that of the general population. Using data from the American Cancer Society’s Study of Cancer Survivors, Zhou et al showed that ovarian cancer survivors had an HRQoL similar to that of the general population 1 year post diagnosis.

 Kunitake et al found that long-term colorectal cancer survivors had better overall physical and mental health compared with the general population.

 Similarly, Greenwald et al found that cervical cancer does not reduce the quality of life of long-term survivors,

 and Schmidt et al found that breast cancer survivors may experience long-term quality of life comparable with that of the general population.

 Finally, Thong et al found that although disease progression reduces cancer survivors’ quality of life, the effect diminishes over time.

Our study echoes the existing literature in finding that long-term cancer survivors fared as well as or sometimes better than the general population. The main contribution of our study is that it investigates short- and long-term cancer survivors’ well-being across a variety of measures in a large, nationally representative sample.

Our findings are important given that recent breakthroughs in cancer treatment have raised the prospect of long-term survival for increased numbers of cancer patients. For example, new immuno-oncology agents, including ipilimumab, pembrolizumab, and nivolumab, have offered significant survival gains to patients,

 and preliminary data suggest the prospect of long-term survival for some metastatic cancer patients.

Our study does have limitations. First, the HRS data do not distinguish between cancer types and, therefore, we were unable to perform separate analyses by tumor type. As such, our results represent the experience of the “average” cancer survivor and will overrepresent those with particularly common or less deadly cancers. 

Second, the most recent nationally representative cohort in the HRS was in 2010, which was a recession year. Because the nationally representative cohort was constructed from the 2010 survey year and all other cohorts were constructed from the 2004-2012 survey years, secular trends could affect the well-being outcomes. However, we controlled for this by including survey year indicators in our analysis.

Third, in studying the well-being of cancer survivors, we necessarily limited the analysis to those who survived. Thus, we cannot definitively determine whether the observed changes in cancer survivors’ well-being between the short- and long-term time horizons are due to cancer survivorship itself or to characteristics potentially common to individuals more likely to survive cancer. 

Finally, we focused mainly on objective measures of well-being rather than subjective measures. However, focusing on objective measures does come with the advantage that, in doing so, we avoided any biases that can result from the way that subjective well-being outcomes are measured, as these outcomes can be subject to self-reporting or measurement bias.

Our study results are striking, given concerns that patients with cancer may experience a low quality of life.

 Although we do find that quality of life is reduced in the short term, we find that it solidly rebounds in the years after diagnosis, becoming comparable with or even better than that of others of similar age and demographic characteristics. The high quality of life experienced by long-term cancer survivors is relevant in the midst of debate about the cost and value of cancer care. Although cost and other up-front considerations are important, it is also important to keep the patient’s perspective in mind. Prior research has shown that patients value therapies that give them a chance of long-term survival.

 Our study shows that the well-being of long-term cancer survivors eventually recovers, becoming similar to or even better than that of the general population. To the extent that long-term survivors enjoy a high quality of life, it makes interventions offering patients with cancer a chance of long-term survival more valuable.

Precision Health Economics (JS, JTS, EvE), Los Angeles, CA; Bristol-Myers Squibb (TO), Plainsboro, NJ; Wake Forest Baptist Medical Center (KB), Winston-Salem, NC; Georgetown University (TD), Washington, DC.

This study compares the well-being of long-term cancer survivors with that of US residents of similar age and demographic characteristics, patients recently diagnosed with cancer, and individuals with chronic illness.

The Well-Being of Long-Term Cancer Survivors

This study examined the relationship between medication adherence, cost sharing measured as out-of-pocket spending, and total annual spending in Medicare beneficiaries with type 2 diabetes (T2D) to evaluate whether pharmacy cost-sharing programs have the potential to decrease adherence. These programs may unintentionally increase the risk of medical complications and may result in higher spending overall.

This retrospective study used 2006 to 2009 Medicare claims data. The sample included patients 65 years or older with T2D (at least 1 claim with International Classification of Diseases, 9th Revision, Clinical Modification codes 250.x0 and 250.x2 and at least 1 antidiabetes drug claim). 

Medication adherence was measured as proportion of days covered over the first 12 months of observation. Spending and adherence outcomes were defined in deciles. 

The sample included 12,305 patient-year observations. Pharmacy spending for patients in the most adherent (10th) decile was 59% higher than that for patients in the least adherent (1st) decile ($4839 vs $3046). Yet, patients in the 10th decile had 49% lower total ($12,531 vs $24,468) and 64% lower medical spending ($7692 vs $21,421) than patients in the 1st decile. Greater out-of-pocket spending was correlated with lower adherence and higher total and medical spending. 

 This study describes a widespread variation in medication adherence, pharmacy cost sharing, and medical spending in a sample of Medicare beneficiaries with T2D. We found that lower adherence was correlated with higher cost sharing in the Medicare population, perhaps because of unobserved confounding factors. However, the existing literature on patients with employer-sponsored insurance suggests some of this correlation may be indicative of causal relationships.

		Increased cost sharing could decrease the likelihood that patients with type 2 diabetes will adhere to their medication and increase diabetes complications and total diabetes spending. 

		Using Medicare claims data (2006-2009), we examined the relationships between medication adherence, cost sharing measured as out-of-pocket spending, and total annual spending to evaluate whether pharmacy cost-sharing programs have the potential to decrease adherence. 

		We found a widespread variation in medication adherence, pharmacy cost sharing, and medical spending, and that lower adherence is correlated with higher cost sharing in the sample of Medicare beneficiaries with type 2 diabetes.

In 2014, nearly 30 million Americans (9.3%) were living with diabetes, including an estimated 11.2 million who were 65 years or older.

 In 2012, the total national spending on diabetes was $245 billion, including $176 billion in direct medical spending. Patients 65 years or older use more healthcare services than younger patients and incur the majority of costs associated with diabetes. Ultimately, the Medicare program—in which 28% of beneficiaries have diabetes&mdash;bears most of this burden.

 Interventions targeted at individuals using the most resources and who have the greatest expenditures have the potential for a disproportionate impact on spending.

Several studies evaluating the economic burden of diabetes have estimated that antidiabetes agents and supplies account for 12% of total healthcare spending.

 Payers have implemented various cost-sharing schemes to reduce pharmacy spending, including incentive-based formularies, coinsurance or co-payments, and benefit caps.

 Research consistently shows that higher co-payments or coinsurance reduce pharmaceutical utilization and adherence among patients with chronic diseases like type 2 diabetes (T2D), heart failure, and schizophrenia,

 and that reducing co-payments in employer-sponsored insurance plans can increase adherence.

Although cost sharing may reduce overall pharmacy spending, it could also have consequences for patient health and result in unintended increases in nonpharmacy healthcare spending, thereby offsetting the savings on pharmaceuticals. The limited existing research shows that cost sharing increases emergency department visits and hospitalizations, especially among chronically ill patients.

 The impact of cost sharing on therapeutic adherence is particularly relevant for patients with diabetes, given that improved adherence is associated with a greater likelihood of reaching glycated hemoglobin (A1C) targets—the primary measure of effective disease management

—and cost sharing has increased since the introduction of Medicare Part D.

 By some estimates, a 1% increase in antidiabetes prescription utilization would reduce medical spending for Medicare beneficiaries with T2D by 0.83%.

Despite the large overarching financial burden of T2D, few studies have examined the extent to which pharmacy and medical spending contribute to total diabetes-related spending.

 This retrospective claims study of a Medicare population examined the relationship between adherence, cost sharing as measured by out-of-pocket (OOP) spending, and total annual spending.

The study used claims data for a 20% random sample of age-eligible Medicare fee-for-service beneficiaries for the years 2006 to 2009. Data were acquired from the CMS Medicare Research Identifiable Files, which provide medical and pharmacy claims, including diagnoses, procedure codes, and measures of charges and spending. Overall spending in any category (total healthcare, medical, or pharmacy) equaled the sum of the patient OOP component and the payer-covered component.

A sample was created using the following criteria: 1) Medicare beneficiaries 65 years or older; and 2) with at least 1 claim with an 

International Classification of Diseases, 9th Revision, Clinical Modification

codes 250.x0 and 250.x2) and at least 1 claim for an antidiabetes medication in the period following the first observed claim with diabetes diagnosis. The index date for patients in the sample was the date of the first observed claim with a T2D diagnosis. Only patients 65 years or older at the time of the first observed claim with a diabetes diagnosis were included. The sample included patients with a T2D diagnosis or antidiabetes medication use in the year prior to the first observed claim in the database if the patient’s index date was less than 1 year after the date of Medicare enrollment. Patients with a follow-up period of less than 1 year after the index date or those who were pregnant or less than 6 months post partum were excluded.

Adherence was measured as the proportion of days covered (PDC), a well-established method that has been used for several decades to evaluate adherence in diabetes and other chronic conditions (eg, cardiovascular disease, dyslipidemia, hypertension).

 PDC was calculated as the total number days when a patient had medication on hand during the observation window (ie, date of first claim to exhaustion of last prescription, segmented into calendar years).

This retrospective study used descriptive analysis to evaluate the extent to which patients’ health status differed across deciles of spending. In particular, we calculated the prevalence of diabetic comorbidities, complications, and drug therapy type by decile of total annual spending. To investigate the relationship between medication adherence and spending, average pharmacy, medical, and total spending were calculated within each adherence decile. To examine the relationship between cost sharing—as measured by OOP spending (total, medical, and pharmacy)&mdash;and medication adherence, average OOP pharmacy, medical, and total spending were calculated within each decile of adherence.

Dividing the sample into deciles created enough bins to characterize the spending distribution. Using deciles to map the distribution of total, pharmaceutical, and medical spending among Medicare beneficiaries with T2D also enabled the identification and description of the degree of skew in annual spending distribution.

Based on the inclusion and exclusion criteria, the sample included 12,305 patients with a mean follow-up period of 3.2 years and a mean age of 74.4 years. The sample was 64% female and 36% male.

Distribution of Total, Medical, and Pharmacy Spending

Considerably higher average total and medical spending were observed for patients in the upper deciles of total spending compared with patients in the lower deciles, but the between-decile difference in pharmacy spending for patients was not as substantial (Figure 1). The average annual total spend for patients in the lowest spending (1st) decile was $794; this increased by nearly 100-fold, to $78,641, for patients in the highest spending (10th) decile. The average annual medical spend for patients in the 1st decile was $417 and increased more than 170-fold, to $71,203, for patients in the 10th decile. However, the average annual pharmacy spend for patients in the 1st decile was $377 compared with $7439 for patients in the 10th decile, a less than 20-fold increase.

Patient Descriptive Statistics and Health Status by Spending

Patient age and sex were similar across all total annual deciles of spending; however, patients in the upper deciles were substantially more likely to have comorbid conditions, including congestive heart failure (74% in the 10th decile vs 17% in the 1st decile), cardiac arrhythmias (68% vs 24%), complicated hypertension (57% vs 17%), or moderate/severe renal impairment (18% vs 1%), among others. The use of certain antihyperglycemic drug therapies was also more prevalent in the upper deciles of total spending. Compared with patients in the 1st annual spending decile, patients with T2D in the 10th decile had a nearly 5-fold higher probability of using long- or rapid-acting insulin to treat their diabetes (31%-34% vs 3%-7%, respectively). Dipeptidyl peptidase-4 inhibitors, meglitinide, and thiazolidinedione use was also higher for patients in the upper deciles of total annual spending, whereas metformin (a biguanide) use was less common in the 10th versus 1st decile (31% vs 53%). Use of sulfonylureas and glucagon-like peptide-1 receptor agonists did not vary significantly across the deciles of total annual spending.

Mean PDC was divided into deciles and ranged from 0.20 for patients in the 1st decile to 1.00 for patients in the 10th decile. As shown in Figure 2, patients with greater medication adherence, as measured by PDC, had lower average annual medical and total spending and only moderately increased pharmacy spending. For patients in the 1st decile of PDC (lowest adherence) to the 10th decile, average annual total spending decreased 49% (from $24,468 to $12,531), and average annual medical spending fell 64% (from $21,421 to $7692). By comparison, pharmacy spending between the 1st decile of PDC and the 10th decile increased 59% (from $3046 to $4839).

An examination of the relationship between total OOP spending and adherence found analogous results. Average annual pharmacy OOP spending for patients in the 1st to 10th PDC decile increased 67% (from $441 to $737), but average annual total OOP spending decreased 36% (from $2116 to $1344), and average annual medical OOP spending decreased 64% (from $1675 to $607).

Based on the present study’s findings, Medicare patients with T2D have highly skewed total annual spending; patients in the upper deciles of total spending account for a disproportionate share of aggregate total spending. As previous studies had similar findings,

 this analysis confirms that medical spending accounts for the vast majority of the skew in total spending on T2D treatment, and that pharmacy spending is a relatively small proportion of total spending.

An analysis of comorbidities and medication use according to total spending showed that patients in the highest deciles were more likely to have comorbid conditions and to use certain antidiabetes drugs (in particular, long-acting, short-acting, and premix insulins). It is likely that the differences by decile in the prevalence of diabetic comorbidities, complications, and therapy utilization in patients play a large role in driving healthcare spending variation. This finding is supported by existing studies.

This study also examined the distribution of total, medical, and pharmacy spending in Medicare patients with T2D based on their level of medication adherence. More adherent patients had substantially lower total spending and medical spending, but moderately increased pharmacy spending, suggesting that, for Medicare patients with T2D, lower adherence is related to increased spending in general. Increased spending may be the result of poorer outcomes, increased comorbidities, and a need for additional treatment, as medication adherence has been significantly associated with improved health status, fewer complications, and less medical resource utilization in patients with T2D.

This study corroborates previous research and suggests that increasing adherence among Medicare beneficiaries with T2D could generate significant cost savings.

 The use of pharmacy cost sharing in Medicare has gone up since the introduction of Medicare Part D, which extended pharmacy benefit coverage and increased pharmaceutical utilization,

 but the availability of generic drugs has prevented significant increases in OOP spending. If this cost-sharing trend continues, adherence may fall and any savings on pharmacy spending could be offset by increased medical spending.

 Recent proposals to limit annual OOP spending for Medicare beneficiaries would reduce the financial burden on high-spend Medicare beneficiaries with T2D, potentially improving medication adherence and health outcomes.

This study has certain limitations. The analysis was descriptive, with data randomly drawn from an administrative claims database representing the years 2006 to 2009. Although the levels of pharmacy and medical spending may have changed since 2009, our qualitative results and the correlations between cost sharing, adherence, and medical spending should not be impaired by the age of the data, as many similar studies have found the same relationships in other patient samples and time periods.

 Additionally, because the study drew from the Medicare database, and because only patients 65 years or older at the time of the first observed T2D diagnosis claim were included, the mean age of the sample (74 years) was higher than the average age of T2D diagnosis (54 years) in the United States.

 Last, only limited medical history is available for patients in claims data research, thereby increasing the potential for confounding, and the date of patients’ initial T2D diagnosis could not be determined with certainty.

This study has several implications. First, the study results demonstrates the existence of a skew in spending on T2D treatment. Medicare patients in the highest deciles of total spending had dramatically higher total and medical spending, clearly indicating that most spending on diabetes-related treatment is from medical—and not pharmacy&mdash;spending among a subset of high-cost patients. In addition, the study expands on available studies of patients with employer-sponsored insurance to demonstrate that improved adherence is related to reduced resource use and spending in Medicare beneficiaries with T2D and suggest that interventions aimed at improving adherence among high-cost patients have the greatest potential to reduce health expenditures. Most importantly, the study results confirm the negative connection established in the existing literature between cost sharing and medication adherence, which varied widely among Medicare beneficiaries with T2D.

 If pharmacy cost-sharing programs decrease medication adherence, initiatives that promote pharmacy cost sharing may unintentionally increase the risk of medical complications offsetting some (or in targeting cases, all) of the savings associated with lower pharmaceutical spending.&ensp;

 Precision Health Economics (JPM, JV-O, JS, DP), Los Angeles, CA; AstraZeneca (JJS, IK), Fort Washington, PA; University of California, Los Angeles (WY), Los Angeles, CA; University of Southern California (ALP), Los Angeles, CA.

This study describes a widespread variation in medication adherence, pharmacy cost sharing, and medical spending. Increased cost sharing may decrease adherence and increase total diabetes spending.

Jeffrey Sullivan, MS

Articles

Evaluating HCV Screening, Linkage to Care, and Treatment Across Insurers

The Well-Being of Long-Term Cancer Survivors

The Relationship Between Adherence and Total Spending Among Medicare Beneficiaries With Type 2 Diabetes

Measuring the Value of Better Diabetes Management