- Email: [email protected]
WILLINGNESS-TO-PAY TREATMENT TIME: PATIENT VERSUS NON-PATIENT PREFERENCES AND THE IMPACT OF ADAPTATION ON MONETARY VALUATION
A270
VA L U E I N H E A LT H 1 9 ( 2 0 1 6 ) A 1 - A 3 1 8
PHP68 A COST-BENEFIT ANALYSIS OF A RAPID, COMPREHENSIVE ORAL NUTRITIONAL SUPPLEMENT QUALITY-IMPROVEMENT PROGRAM FOR MALNOURISHED HOSPITALIZED PATIENTS Sulo S 1, Goates S 2, Partridge J 2, Feldstein J 3, Summerfelt W T 1, Hegazi R 2, Sriram K 1 1Advocate Health Care, Downers Grove, IL, USA, 2Abbott Nutrition, Columbus, OH, USA, 3Center for Applied Value Analysis, Great Barrington, MA, USA .
.
.
.
.
.
.
.
Objectives: To conduct a cost-benefit analysis of a quality improvement program (QIP) for malnourished hospitalized patients. Methods: The QIP utilized a prospective, two-group, pre-post, multi-site design. Historical controls were utilized to determine baseline 30-day readmission rates for malnourished patients receiving oral nutritional supplements (ONS). These rates were compared to 30-day readmission rates of two prospective cohorts (“QIP” and “QIP+”) in 4 hospitals of a large system. In 2 QIP hospitals (n= 769), the Electronic Medical Record (EMR) was upgraded to include a validated malnutrition screening tool and automatic condition-specific administration of ONS for all patients at risk for malnutrition. Two QIP+ hospitals (n= 500) implemented the same improvements as the QIP group along with faster delivery of ONS, ONS patient education at discharge, and postdischarge compliance calls. Results: Both groups achieved a statistically significant reduction in 30-day readmission compared to historical controls: 22% for QIP+ (p< 0.01) and 18% for QIP (p< 0.01). The costs of the QIP+ and QIP programs were $49,564 and $40,142, respectively. Fixed costs included hospital administration; information technology; QIP management; and healthcare provider education. Variable costs included post-discharge calls and patient screening/enrollment. QIP+ savings was based on a 22% reduction in 30-day readmission rates where baseline readmission was estimated to be 100 (20% of 500) and was reduced to 78 (15.6% of 500). Using a readmission cost of $18,478 (Philipson et al., 2013), a savings of $406,516 from avoided readmissions was noted; when subtracting the $49,564 program costs, a $714 per patient net savings was achieved. Likewise, the QIP group realized a $511,545 savings from avoided readmissions, and a $613 per patient net savings. Conclusions: A rapid, comprehensive ONS QIP was found to reduce 30-day readmissions among malnourished hospitalized patients resulting in improved patient outcomes and reduced costs. PHP69 HOSPITAL STAFF PHARMACIST PRODUCTIVITY: A COMPARATIVE ANALYSIS OF TIME SPENT BY PHARMACISTS IN A COMPUTERIZED PROVIDER ORDER ENTRY (CPOE) SYSTEM VERSUS NON-CPOE SYSTEM Lewing B D 1, Hatfield M D 1, Sansgiry S S 2 1University of Houston, Houston, TX, USA, 2The University of Houston, Houston, TX, USA .
.
.
.
.
.
Objectives: The objective of this study was to measure time spent, based on type of activity, by hospital staff pharmacists in a setting utilizing computerized provider order entry (CPOE) and a setting in which CPOE was not utilized. Then, using the data collected, complete a comparative analysis of productivity and time usage between CPOE and non-CPOE settings. Methods: Staff pharmacists were observed in two community teaching hospitals within the same healthcare system in Houston, TX – one CPOE (252 beds) and one non-CPOE (274 beds). A pre-validated instrument was used to record 37 different pharmacist tasks which were sorted into six different activity categories. Time spent by pharmacists on different activity categories were compared between the CPOE and non-CPOE settings. The Wilcoxon Two-Sample Test was used to test the significance of difference between the two samples for each of the six activity categories. Data analysis was completed using SAS version 9.3, with significance set at 0.05. Results: A total of 77 hours of data were collected for each the CPOE pharmacy setting and for the non-CPOE pharmacy setting. The amount of time spent by pharmacists at the different settings were (mean number of minutes per hour ± SD CPOE, non-CPOE, p-value): order entry (17.6 ± 10.7, 29.6 ± 11.3, p< 0.05); order verification (14.0 ± 8.6, 0.9 ± 1.8, p< 0.05); distributive tasks (16.0 ± 8.3, 13.8 ± 10.0, p< 0.05); clinical (5.0 ± 4.2, 6.6± 6.4, p= 0.41); administrative (5.4 ± 5.9, 5.6 ± 6.8, p= 0.83); and miscellaneous (2.0 ± 2.9, 3.6 ± 4.3, p< 0.05). Conclusions: At the CPOE pharmacy less time was spent on order entry and more time was spent on order verification and distributive tasks. The presence of a CPOE system could have a dramatic effect on pharmacist workflow productivity and time spent on certain activities. PHP70 PROJECTING THE BURDEN OF CHRONIC DISEASES IN THE UNITED STATES FROM 2016 TO 2025 UNDER DIFFERENT SCENARIOS USING A MICROSIMULATION APPROACH Su W , Chen F , Storm M , Semilla A , Iacobucci W , Dall T IHS Life Sciences, Washington, DC, USA .
.
.
.
.
.
Objectives: Chronic diseases impose significant financial and health burdens on the society. The aim of this study is to project the health and financial burdens of chronic diseases from societal and employer perspectives until 2025, and estimate the potential benefits of population health initiatives and medical advancements. Methods: The baseline population was synthesized from large national survey databases including National Health and Nutrition Examination Survey (NHANES), American Community Survey (ACS), Behavioral Risk Factor Surveillance System (BRFSS), and National Nursing Home Survey (NNHS) using a propensity matching algorithm. A microsimulation model was constructed in SAS to simulate the prevalence, incidence, progression, and potential treatment effects of more than 10 chronic diseases including asthma, COPD, osteoporosis, cancers, diabetes, hypertension, stroke, and other cardiovascular diseases. Each individual’s demographics, biometrics, diagnosis and treatment status, and health history were used to predict disease incidence, progression, and mortality based on published clinical evidence. The baseline scenario assumes current population behavior and treatment states continue into the future. The ‘optimistic’ scenario assumes more people quit smoking and drinking, more physically active population, higher treatment adherence, timely diagnosis, slower cost growth, and insurance coverage expansion. Results: If current trends continue, chronic diseases will result in 1.4 trillion in direct medi-
cal costs and 802 million work days in absenteeism in 2025. In the same year 210 million people will have at least 1 chronic disease. Treating the chronic conditions will consume 35 cents out of every dollar spent on healthcare in the following 10 years. The ‘optimistic’ scenario would result in 51 million fewer diagnoses and 780 billion in total direct medical cost savings by 2025. Conclusions: Costs associated with chronic diseases constitute a significant portion of all healthcare spending. Investing in population health initiatives could effectively reduce the burden 10 years from now. PHP71 WILLINGNESS-TO-PAY TREATMENT TIME: PATIENT VERSUS NON-PATIENT PREFERENCES AND THE IMPACT OF ADAPTATION ON MONETARY VALUATION Bobinac A Erasmus University Rotterdam, Rotterdam, The Netherlands .
Objectives: We conducted a Contingent valuation study to estimate the monetary value (WTP) of a unit of treatment time in the domain of hemodialysis treatment in three distinct samples: (1) actual hemodialysis patients; (2) pre-dialysis patients suffering chronic renal disease (CRD) stage 3 and 4; (3) healthy controls. Next reporting the estimates of a unit of treatment time, we analysed if patient’s adaptation to dialysis effects of the value of treatment time. Methods: The data was collected in the University hospital in Rijeka, Croatia. The data included the information regarding (1) the proportion of the actual and hypothetical patients who prefer to reduce their treatment time spent on dialysis (from 4 to 1 hour); (2) the proportion of the actual and hypothetical patients who are willing to pay out of pocket for the reduction; (3) the average WTP per unit of time in the three groups. Results: Highest WTP for hour of treatment time (€ 3.42) was reported by healthy controls. Pre-dialysis patient estimated the value of € 1.18 while dialysis patients reported the value of € 0.91 per hour of treatment time. We found a significant and inversely correlated relationship between the monetary value of treatment time and the length of time spent receiving dialysis treatment in the group of dialysis patients. To test whether adaptation is in fact a consequence of a response shift, or a change in patient’s reference point, we compared the valuations of the same health state obtained from the three subject groups. Patients and non-patient provided the same VAS scores for the same hypothetical condition suggesting no response shift. Conclusions: The value of treatment time should not be ignored in economic evaluations. WTP and stated preference methods more generally should account for the patients’ adaptation to stable states. The development of systematic recommendations is warranted. PHP72 DOWNSTREAM MEDICAL COST OFFSETS IN MEDICARE FROM PRESCRIPTION DRUG USE IN EMPLOYER-SPONSORED INSURANCE Roebuck M C RxEconomics LLC, Hunt Valley, MD, USA .
.
Objectives: To estimate the impact of prescription drug utilization during preMedicare coverage years under employer-sponsored insurance on later medical spending in Medicare. Methods: Individuals found in both the Commercial and Medicare Supplemental Marketscan® files were selected. Study subjects were required to be 64 years old or younger at baseline and continuously eligible throughout the entire 5-year study period 2008-2012. The resulting panel dataset was comprised of 133,412 individuals who aged into Medicare, each with 5 annual observations. Conducted was an econometric analysis of the impact of aggregate prescription drug utilization on total, inpatient, and outpatient medical costs. Individual fixed effects Poisson models of these cost measures were estimated as a function of 1-, 2-, and 3-year lagged values of prescription drug utilization—measured as the number of days’ supply-adjusted fills. All models also included a robust set of controls. Results: Among individuals covered by Medicare (ages 65-68), a 1 percent increase in prescription drug utilization 1 year prior was associated with a 0.363 percent decrease in total medical costs. Notably, medication use 2 years prior prompted a 0.123 percent decrease in total medical costs, and these offsets still persisted after 3 years (0.059 percent decrease). Effects were more pronounced for inpatient relative to outpatient costs. The pattern of declining, yet persistent, medical cost offsets for at least 3 years emerged for all three measures of medical costs. Conclusions: Pharmaceutical use is expected to lower spending on other medical services. The duration of these offsets may span multiple payers. Retirees entering Medicare who are healthier due to prior medication use are less costly. Indeed, the Medicare program is exposed to the consequences of prescription drug use 3 years prior to coverage. The evaluation of policies that could alter medication use among the near retired (such as the “Cadillac Tax”) should consider this important finding. PHP73 COST OF CHRONIC CONDITIONS FOR EMPLOYEES AND EMPLOYERS IN THE UNITED STATES Ghushchyan V , Kolian S , Manukyan S American University of Armenia, Yerevan, Armenia .
.
.
Objectives: The average age of working population in the US continues rising while the average age of some prevalent chronic conditions keeps decreasing. With these changing demographics the evaluation of indirect (productivity) cost associated with chronic conditions becomes actual for both employees and employers. The aim of this paper is to estimate the cost associated with chronic conditions for employees and employers in the US. Methods: Pooled 2011-2012 Medical Expenditure Panel Survey data was used to estimate the number of days an average US working age employee is missing work due to chronic condition. The missed work-days were calculated for 1,2,…7 and 8+ chronic conditions. A negative binomial regression was used to regress missed work-days on dummy variables for 1 to 8+ chronic conditions controlling for socio-demographic characteristics. Estimated regression-adjusted number of missed work-days associated with the number of chronic conditions were multiplied by the average US daily working hours and mean wages to derive per employee lost wages. In addition indirect cost, that is supervisors’ and other
VA L U E I N H E A LT H 1 9 ( 2 0 1 6 ) A 1 - A 3 1 8
PHP68 A COST-BENEFIT ANALYSIS OF A RAPID, COMPREHENSIVE ORAL NUTRITIONAL SUPPLEMENT QUALITY-IMPROVEMENT PROGRAM FOR MALNOURISHED HOSPITALIZED PATIENTS Sulo S 1, Goates S 2, Partridge J 2, Feldstein J 3, Summerfelt W T 1, Hegazi R 2, Sriram K 1 1Advocate Health Care, Downers Grove, IL, USA, 2Abbott Nutrition, Columbus, OH, USA, 3Center for Applied Value Analysis, Great Barrington, MA, USA .
.
.
.
.
.
.
.
Objectives: To conduct a cost-benefit analysis of a quality improvement program (QIP) for malnourished hospitalized patients. Methods: The QIP utilized a prospective, two-group, pre-post, multi-site design. Historical controls were utilized to determine baseline 30-day readmission rates for malnourished patients receiving oral nutritional supplements (ONS). These rates were compared to 30-day readmission rates of two prospective cohorts (“QIP” and “QIP+”) in 4 hospitals of a large system. In 2 QIP hospitals (n= 769), the Electronic Medical Record (EMR) was upgraded to include a validated malnutrition screening tool and automatic condition-specific administration of ONS for all patients at risk for malnutrition. Two QIP+ hospitals (n= 500) implemented the same improvements as the QIP group along with faster delivery of ONS, ONS patient education at discharge, and postdischarge compliance calls. Results: Both groups achieved a statistically significant reduction in 30-day readmission compared to historical controls: 22% for QIP+ (p< 0.01) and 18% for QIP (p< 0.01). The costs of the QIP+ and QIP programs were $49,564 and $40,142, respectively. Fixed costs included hospital administration; information technology; QIP management; and healthcare provider education. Variable costs included post-discharge calls and patient screening/enrollment. QIP+ savings was based on a 22% reduction in 30-day readmission rates where baseline readmission was estimated to be 100 (20% of 500) and was reduced to 78 (15.6% of 500). Using a readmission cost of $18,478 (Philipson et al., 2013), a savings of $406,516 from avoided readmissions was noted; when subtracting the $49,564 program costs, a $714 per patient net savings was achieved. Likewise, the QIP group realized a $511,545 savings from avoided readmissions, and a $613 per patient net savings. Conclusions: A rapid, comprehensive ONS QIP was found to reduce 30-day readmissions among malnourished hospitalized patients resulting in improved patient outcomes and reduced costs. PHP69 HOSPITAL STAFF PHARMACIST PRODUCTIVITY: A COMPARATIVE ANALYSIS OF TIME SPENT BY PHARMACISTS IN A COMPUTERIZED PROVIDER ORDER ENTRY (CPOE) SYSTEM VERSUS NON-CPOE SYSTEM Lewing B D 1, Hatfield M D 1, Sansgiry S S 2 1University of Houston, Houston, TX, USA, 2The University of Houston, Houston, TX, USA .
.
.
.
.
.
Objectives: The objective of this study was to measure time spent, based on type of activity, by hospital staff pharmacists in a setting utilizing computerized provider order entry (CPOE) and a setting in which CPOE was not utilized. Then, using the data collected, complete a comparative analysis of productivity and time usage between CPOE and non-CPOE settings. Methods: Staff pharmacists were observed in two community teaching hospitals within the same healthcare system in Houston, TX – one CPOE (252 beds) and one non-CPOE (274 beds). A pre-validated instrument was used to record 37 different pharmacist tasks which were sorted into six different activity categories. Time spent by pharmacists on different activity categories were compared between the CPOE and non-CPOE settings. The Wilcoxon Two-Sample Test was used to test the significance of difference between the two samples for each of the six activity categories. Data analysis was completed using SAS version 9.3, with significance set at 0.05. Results: A total of 77 hours of data were collected for each the CPOE pharmacy setting and for the non-CPOE pharmacy setting. The amount of time spent by pharmacists at the different settings were (mean number of minutes per hour ± SD CPOE, non-CPOE, p-value): order entry (17.6 ± 10.7, 29.6 ± 11.3, p< 0.05); order verification (14.0 ± 8.6, 0.9 ± 1.8, p< 0.05); distributive tasks (16.0 ± 8.3, 13.8 ± 10.0, p< 0.05); clinical (5.0 ± 4.2, 6.6± 6.4, p= 0.41); administrative (5.4 ± 5.9, 5.6 ± 6.8, p= 0.83); and miscellaneous (2.0 ± 2.9, 3.6 ± 4.3, p< 0.05). Conclusions: At the CPOE pharmacy less time was spent on order entry and more time was spent on order verification and distributive tasks. The presence of a CPOE system could have a dramatic effect on pharmacist workflow productivity and time spent on certain activities. PHP70 PROJECTING THE BURDEN OF CHRONIC DISEASES IN THE UNITED STATES FROM 2016 TO 2025 UNDER DIFFERENT SCENARIOS USING A MICROSIMULATION APPROACH Su W , Chen F , Storm M , Semilla A , Iacobucci W , Dall T IHS Life Sciences, Washington, DC, USA .
.
.
.
.
.
Objectives: Chronic diseases impose significant financial and health burdens on the society. The aim of this study is to project the health and financial burdens of chronic diseases from societal and employer perspectives until 2025, and estimate the potential benefits of population health initiatives and medical advancements. Methods: The baseline population was synthesized from large national survey databases including National Health and Nutrition Examination Survey (NHANES), American Community Survey (ACS), Behavioral Risk Factor Surveillance System (BRFSS), and National Nursing Home Survey (NNHS) using a propensity matching algorithm. A microsimulation model was constructed in SAS to simulate the prevalence, incidence, progression, and potential treatment effects of more than 10 chronic diseases including asthma, COPD, osteoporosis, cancers, diabetes, hypertension, stroke, and other cardiovascular diseases. Each individual’s demographics, biometrics, diagnosis and treatment status, and health history were used to predict disease incidence, progression, and mortality based on published clinical evidence. The baseline scenario assumes current population behavior and treatment states continue into the future. The ‘optimistic’ scenario assumes more people quit smoking and drinking, more physically active population, higher treatment adherence, timely diagnosis, slower cost growth, and insurance coverage expansion. Results: If current trends continue, chronic diseases will result in 1.4 trillion in direct medi-
cal costs and 802 million work days in absenteeism in 2025. In the same year 210 million people will have at least 1 chronic disease. Treating the chronic conditions will consume 35 cents out of every dollar spent on healthcare in the following 10 years. The ‘optimistic’ scenario would result in 51 million fewer diagnoses and 780 billion in total direct medical cost savings by 2025. Conclusions: Costs associated with chronic diseases constitute a significant portion of all healthcare spending. Investing in population health initiatives could effectively reduce the burden 10 years from now. PHP71 WILLINGNESS-TO-PAY TREATMENT TIME: PATIENT VERSUS NON-PATIENT PREFERENCES AND THE IMPACT OF ADAPTATION ON MONETARY VALUATION Bobinac A Erasmus University Rotterdam, Rotterdam, The Netherlands .
Objectives: We conducted a Contingent valuation study to estimate the monetary value (WTP) of a unit of treatment time in the domain of hemodialysis treatment in three distinct samples: (1) actual hemodialysis patients; (2) pre-dialysis patients suffering chronic renal disease (CRD) stage 3 and 4; (3) healthy controls. Next reporting the estimates of a unit of treatment time, we analysed if patient’s adaptation to dialysis effects of the value of treatment time. Methods: The data was collected in the University hospital in Rijeka, Croatia. The data included the information regarding (1) the proportion of the actual and hypothetical patients who prefer to reduce their treatment time spent on dialysis (from 4 to 1 hour); (2) the proportion of the actual and hypothetical patients who are willing to pay out of pocket for the reduction; (3) the average WTP per unit of time in the three groups. Results: Highest WTP for hour of treatment time (€ 3.42) was reported by healthy controls. Pre-dialysis patient estimated the value of € 1.18 while dialysis patients reported the value of € 0.91 per hour of treatment time. We found a significant and inversely correlated relationship between the monetary value of treatment time and the length of time spent receiving dialysis treatment in the group of dialysis patients. To test whether adaptation is in fact a consequence of a response shift, or a change in patient’s reference point, we compared the valuations of the same health state obtained from the three subject groups. Patients and non-patient provided the same VAS scores for the same hypothetical condition suggesting no response shift. Conclusions: The value of treatment time should not be ignored in economic evaluations. WTP and stated preference methods more generally should account for the patients’ adaptation to stable states. The development of systematic recommendations is warranted. PHP72 DOWNSTREAM MEDICAL COST OFFSETS IN MEDICARE FROM PRESCRIPTION DRUG USE IN EMPLOYER-SPONSORED INSURANCE Roebuck M C RxEconomics LLC, Hunt Valley, MD, USA .
.
Objectives: To estimate the impact of prescription drug utilization during preMedicare coverage years under employer-sponsored insurance on later medical spending in Medicare. Methods: Individuals found in both the Commercial and Medicare Supplemental Marketscan® files were selected. Study subjects were required to be 64 years old or younger at baseline and continuously eligible throughout the entire 5-year study period 2008-2012. The resulting panel dataset was comprised of 133,412 individuals who aged into Medicare, each with 5 annual observations. Conducted was an econometric analysis of the impact of aggregate prescription drug utilization on total, inpatient, and outpatient medical costs. Individual fixed effects Poisson models of these cost measures were estimated as a function of 1-, 2-, and 3-year lagged values of prescription drug utilization—measured as the number of days’ supply-adjusted fills. All models also included a robust set of controls. Results: Among individuals covered by Medicare (ages 65-68), a 1 percent increase in prescription drug utilization 1 year prior was associated with a 0.363 percent decrease in total medical costs. Notably, medication use 2 years prior prompted a 0.123 percent decrease in total medical costs, and these offsets still persisted after 3 years (0.059 percent decrease). Effects were more pronounced for inpatient relative to outpatient costs. The pattern of declining, yet persistent, medical cost offsets for at least 3 years emerged for all three measures of medical costs. Conclusions: Pharmaceutical use is expected to lower spending on other medical services. The duration of these offsets may span multiple payers. Retirees entering Medicare who are healthier due to prior medication use are less costly. Indeed, the Medicare program is exposed to the consequences of prescription drug use 3 years prior to coverage. The evaluation of policies that could alter medication use among the near retired (such as the “Cadillac Tax”) should consider this important finding. PHP73 COST OF CHRONIC CONDITIONS FOR EMPLOYEES AND EMPLOYERS IN THE UNITED STATES Ghushchyan V , Kolian S , Manukyan S American University of Armenia, Yerevan, Armenia .
.
.
Objectives: The average age of working population in the US continues rising while the average age of some prevalent chronic conditions keeps decreasing. With these changing demographics the evaluation of indirect (productivity) cost associated with chronic conditions becomes actual for both employees and employers. The aim of this paper is to estimate the cost associated with chronic conditions for employees and employers in the US. Methods: Pooled 2011-2012 Medical Expenditure Panel Survey data was used to estimate the number of days an average US working age employee is missing work due to chronic condition. The missed work-days were calculated for 1,2,…7 and 8+ chronic conditions. A negative binomial regression was used to regress missed work-days on dummy variables for 1 to 8+ chronic conditions controlling for socio-demographic characteristics. Estimated regression-adjusted number of missed work-days associated with the number of chronic conditions were multiplied by the average US daily working hours and mean wages to derive per employee lost wages. In addition indirect cost, that is supervisors’ and other