Advertisement

Cost of Human Immunodeficiency Virus (HIV) and Determinants of Healthcare Costs in HIV-Infected Treatment-Naive Patients Initiated on Antiretroviral Therapy in Germany: Experiences of the PROPHET Study

Open ArchivePublished:August 14, 2020DOI:https://doi.org/10.1016/j.jval.2020.04.1836

      Highlights

      The presented findings not only confirm existing (national and international) knowledge about components and driving factors of healthcare costs in human immunodeficiency virus infected persons, but also provide new insights. Current cost information has been determined that differ from those found in older studies, especially owing to changes in antiretroviral therapy (ART) prices, which are the driving cost factor. In addition, because only patients initiating ART were included, the changes in costs within the first 2 years of ART could be observed.

      Abstract

      Objectives

      The purpose of the prospective clinical and pharmacoeconomic outcomes study of different first-line antiretroviral treatment strategies (PROPHET) was to examine the healthcare costs of human immunodeficiency virus (HIV)–infected persons in Germany treated with different antiretroviral therapy (ART) strategies and to identify variables associated with high costs.

      Methods

      The setting was a 24-month prospective multicenter observational cohort study in a German HIV-specialized care setting from 2014 to 2017. A microcosting approach was used for the estimation of healthcare costs. Data were obtained via electronic case report forms. The costs were calculated from both the societal and the statutory health insurance perspective. Regression models were performed that took into consideration the impact of several independent variables.

      Results

      Four hundred thirty-four patients from 24 centers throughout Germany were included. Average annual healthcare costs were €20 118 (standard deviation [SD] €6451) per patient from the societal perspective (n = 336) and €17 306 (SD €4106) from the statutory health insurance perspective (n = 292). Expenditures for the ART medication had the highest impact. Total costs declined in the second year of therapy. There was a significant association between the amount of total cost and clinical or therapeutic variables from both perspectives; a diagnosis of acquired immune deficiency syndrome (AIDS) led to higher costs as well as the chosen ART strategy. Age also increased cost from the statutory health insurance perspective.

      Conclusions

      The main cost driver of the healthcare costs for HIV-positive patients was antiretroviral drug expenses. Further variables that influenced the costs were identified. The results provide a detailed overview of the resource use of patients in the PROPHET cohort.

      Introduction

      After a sharp decline at the end of the 1990s, the rate of new human immunodeficiency virus (HIV) infections in Germany has stabilized at an elevated level since 2006 and has been at approximately 3200 persons per year since 2014. Currently, it is estimated that over 87 900 people live with human immunodeficiency virus or acquired immune deficiency syndrome in Germany.
      • an der Heiden M.
      • Marcus U.
      • Kollan C.
      • et al.
      Schätzung der Zahl der HIV-Neuinfektionen und der Gesamtzahl von Menschen mit HIV in Deutschland.
      HIV infection and AIDS pose a significant economic burden, especially because it is a chronic disease requiring lifelong therapy. Previously published data estimate a cost of more than €22 000 per patient per year from a societal perspective in Germany, with slightly lower costs from the statutory health insurance perspective. The cost of antiretroviral therapy (ART) makes up the largest cost component and increases with disease progression.
      • Mostardt S.
      • Hanhoff N.
      • Wasem J.
      • et al.
      Cost of HIV and determinants of health care costs in HIV-positive patients in Germany: results of the DAGNÄ K3A Study.
      ,
      • Treskova M.
      • Kuhlmann A.
      • Bogner J.
      • et al.
      Analysis of contemporary HIV/AIDS health care costs in Germany: driving factors and distribution across antiretroviral therapy lines.
      Ninety-three percent of persons with known HIV infection received ART in 2018.
      • an der Heiden M.
      • Marcus U.
      • Kollan C.
      • et al.
      Schätzung der Zahl der HIV-Neuinfektionen und der Gesamtzahl von Menschen mit HIV in Deutschland.
      Early diagnosis and initiation of treatment can result in high levels of virologic control, a good immune reconstitution and a low morbidity subsequently enabling the HIV-infected individuals to lead a healthy life without reductions in productivity. In addition, high rates of successfully treated patients reduce the spread of HIV in a given population.
      US Department of Health and Human Services
      Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in Adults and Adolescents Living with HIV.
      • Rodger A.J.
      • Cambiano V.
      • Bruun T.
      • et al.
      Sexual activity without condoms and risk of HIV transmission in serodifferent couples when the HIV-positive partner is using suppressive antiretroviral therapy.
      • Cohen M.S.
      • Chen Y.Q.
      • McCauley M.
      • et al.
      Antiretroviral therapy for the prevention of HIV-1 transmission.
      Over 27 antiretroviral agents have been licensed and are used in different combinations with various intake modalities, potential adverse effects, resistance profiles, and interactions. It is expected that different combinations of antiretroviral agents used for treatment are also associated with differences in the cost of illness.
      • Stoll M.
      • Kollan C.
      • Bergmann F.
      • et al.
      Calculation of direct antiretroviral treatment costs and potential cost savings by using generics in the German HIV ClinSurv cohort.
      The objective of the prospective clinical and pharmacoeconomic outcomes study of different first-line antiretroviral treatment strategies (PROPHET) was to evaluate the pharmacoeconomic and clinical outcomes of different first-line antiretroviral treatment regimens recommended by treatment guidelines in Germany. The aim of the pharmacoeconomic part of the study was to determine the cost associated with various first-line combination ART strategies and to investigate any differences between them. Clinical data are presented elsewhere.
      • Bickel M.
      • Hoffmann C.
      • Wolf E.
      • et al.
      High effectiveness of recommended first-line antiretroviral therapies in Germany: a nationwide, prospective cohort study.
      In this report, we present the results of the pharmacoeconomic analysis of the PROPHET study, which may be relevant on different levels of the healthcare system. Aside from insights for health professionals, PROPHET offers relevant information on costs and economic impact for decision makers and scientists (for example, for international comparisons or evaluations of the HIV pre-exposure prophylaxis).

      Methods

      Study Design

      PROPHET was conducted by the German Association of Physicians specializing in HIV care registered association. (DAGNÄ e.V.) in cooperation with the Institute for Health Care Management and Research of the University of Duisburg-Essen, MUC Research GmbH, and Clinovate NET GmbH & Co KG. The PROPHET study is a prospective, nationwide, multicenter health economic and clinical evaluation with a societal and health insurance perspective. The study was approved by the ethics committee of the medical department of the University of Duisburg-Essen and, where required, by the local ethics committees of the participating centers. The analysis of health economic aspects was carried out by the Institute for Health Care Management and Research of the University of Duisburg-Essen. The observation period was 24 months for every patient.

      Setting

      Private practices and ambulatory care centers specializing in HIV care throughout Germany were invited to take part in this study. This reflects the usual care of patients with HIV in Germany. Patients were eligible for study participation if they (1) were HIV-positive, (2) initiated an ART (baseline = start of the ART) as recommended by current guidelines (containing 2 nucleoside/nucleotide reverse transcriptase inhibitors [NRTI] plus a third agent, either an integrase strand transfer inhibitor [INSTI], a boosted protease inhibitor [PI] or a non-nucleoside reverse transcriptase inhibitor [NNRTI]), (3) were at least 18 years of age, and (4) provided written consent. Additional criteria for inclusion into the health economic analysis were: at least 1 completed health economic report form at each study visit; no loss to follow-up; and insured by statutory health insurance at each study visit for the statutory health insurance perspective.

      Data Collection

      Data were collected at baseline and at 3, 6, 12, 18, and 24 months. The first patient was recruited in August 2014, the last one in September 2015. Data collection and preparation were completed in August 2018. Clinical data were obtained using electronic case report forms and included information about disease status (late presentation, US Centers for Disease Control and Prevention stage, CD4+ T cell count, AIDS-defining illnesses), use of ART, and other medication (categorized). Health economic forms asked for occupation, health insurance status, nursing care, reduction in or loss of earning capacity (form 1), outpatient physician contacts (excluding HIV-specialized physicians) (form 2), hospitalization, inpatient and outpatient rehabilitation, disability (form 3), and detailed information about contacts to HIV-specialized physicians (form 4).

      Cost Calculations

      The health economic part of the PROPHET study included direct medical and indirect costs (productivity loss owing to a reduction in or loss of earning capacity or disability) in HIV-infected patients. Intangible costs were not considered in this evaluation. Healthcare costs were calculated from a societal perspective and from a statutory health insurance perspective, following a bottom-up approach. All costs are reported in 2017 German prices in euros. If necessary, cost data were adjusted by the harmonized German consumer price index. Costs were determined from different sources for both perspectives. The average selling prices for the German market according to the Lauer Taxe were used for ART cost calculation. Costing for other medication followed the same approach, using the average price of the cheapest three N3 packages (the largest standardized package size in Germany) in every category. Depending on the perspective, a manufacturer discount, a statutory health insurance discount, and patient co-payment were considered. In a main scenario, the ART costs were calculated according to the respective study arm without consideration of potential changes in ART strategy. In an alternative scenario, all realized regime changes were taken into account.
      Costing for hospitalizations secondary to HIV infection used the point values of the German Diagnosis Related Groups of the Major Diagnostic category “18A HIV” weighted with their relative frequency in 2017 and the average state base rate (Landesbasisfallwert) in Germany in 2017.
      AOK Bundesverband
      Übersicht über die für 2017 gültigen Landesbasisfallwerte in den einzelnen Bundesländern. 2017.
      ,
      Institut für das Entgeltsystem im Krankenhaus
      G-DRG-Report-Browser 2017.
      For the societal perspective, these costs were supplemented with daily investment costs according to Bock et al.
      • Bock J.-O.
      • Brettschneider C.
      • Seidl H.
      • et al.
      Ermittlung standardisierter Bewertungssätze aus gesellschaftlicher Perspektive für die gesundheitsökonomische Evaluation.
      Hospital contacts due to comorbidities and contacts with physicians in the ambulatory setting (excluding HIV-specialized physicians) due to HIV infection or comorbidities were valued using the 2011 resource values from Bock et al.
      • Bock J.-O.
      • Brettschneider C.
      • Seidl H.
      • et al.
      Ermittlung standardisierter Bewertungssätze aus gesellschaftlicher Perspektive für die gesundheitsökonomische Evaluation.
      Costs of contacts with HIV-specialized physicians were calculated using standard values from the payment system of German physicians. Detailed information from the accounting system of participating HIV-specialized physicians was used for HIV-infected persons insured by private insurance. Rehabilitation costs were calculated similar to the approach of Bock et al using data from the statutory health insurance and statutory pension insurance from 2016.
      Bundesministerium für Gesundheit
      Gesetzliche Krankenversicherung. Vorläufige Rechnungsergebnisse 1.-4. Quartal 2016. 2017.
      Bundesministerium für Gesundheit
      Gesetzliche Krankenversicherung. Endgültige Rechnungsergebnisse 2016. 2017.
      Bundesministerium für Gesundheit
      Ergebnisse der Statistik KG 5, Vorsorge- und Rehabilitationsmaßnahmen 2016 der Gesetzlichen Krankenversicherungen. 2017.
      Deutsche Rentenversicherung Bund
      Reha-Bericht 2018. 2018.
      To compute the nursing care costs, the documented levels of care were adapted into the new classification (since 2017) according to the §140 German Code of Social Law XI (SGB XI). Payments of the German statutory long-term care insurance were taken into consideration. These payments were supplemented by the average additional out-of-pocket costs of the beneficiary.
      Bundesministerium für Gesundheit
      Abschlussbericht zur Studie ‘Wirkungen des Pflege-Weiterentwicklungsgesetzes:’ Bericht zu den Repräsentativerhebungen im Auftrag des Bundesministeriums für Gesundheit. 2011.
      ,
      Deutscher Bundestag
      Antwort der Bundesregierung auf die Kleine Anfrage der Abgeordneten Elisabeth Scharfenberg, Maria Klein-Schmeink, Kordula Schulz-Asche, weiterer Abgeordneter und der Fraktion BÜNDNIS 90/DIE GRÜNEN–Drucksache 18/13453–Umsetzung der Pflegestärkungsgesetze der Bundesregierung. 2017.
      With regard to indirect costs for the statutory health insurance perspective, cost of disability of more than 42 days was computed as the sum of the average statutory health insurance expense for sick pay and the estimated lost insurance premiums per day. Calculation of the daily costs owing to reduction in or loss of earning capacity and disability from the societal perspective were based on average labor costs. These costs were calculated according to the friction approach, which considers that loss of productivity will be compensated, with an assumption of 80% productivity loss per day over a friction period of 90 days.
      • Koopmanschap M.A.
      • Rutten F.F.
      • van Ineveld B.M.
      • et al.
      The friction cost method for measuring indirect costs of disease.

      Analytical Framework

      Costs for single categories and total costs were computed as the product of estimated prices and the amount of consumed resources. Frequencies, mean values, and standard deviations of resource use and healthcare costs were determined (IBM SPSS Statistics version 25). Additionally, healthcare costs of the different study arms were analyzed for statistically significant differences using the Mann-Whitney U test. Furthermore, the Mann-Whitney U test (age, CD4 cell count) and the binominal test (sex, late presentation, AIDS, mode of transmission) were used to search for differences in patient characteristics between the NRTI/PI study arm and the most frequent ART strategy (INSTI arm). To determine which variables significantly influence total cost, various linear regression models were employed. In 2 models (1 for each perspective), the annual total healthcare cost was used as the dependent variable. Independent variables in all models were age at baseline, sex, late presentation, presence of AIDS at baseline, mode of transmission, and ART strategy. The most frequent characteristics in the PROPHET cohort were used as reference categories.
      • Backhaus K.
      • Erichson B.
      • Plinke W.
      • et al.
      Multivariate Analysemethoden. Eine anwendungsorientierte Einführung.
      The reference patient was defined as being male, without late presentation or AIDS, infected by sexual contact between men who have sex with men (MSM), and receiving ART with an INSTI as the third agent. Nonparametric bootstrapping was utilized to assign measures of accuracy (confidence interval [CI]). Variance inflation factors (VIFs) were chosen to check for potential multicollinearity (threshold for assumed multicollinearity, VIF > 10).
      • Greene W.H.
      Econometric analysis.

      Results

      In total 336 of 434 patients met the inclusion criteria for the health economic analysis from the societal perspective. Baseline data of the cohort are shown in Table 1. As shown, patient characteristics differed among the study arms. Significant differences were found between the INSTI and the NNRTI arms and between the INSTI and the PI arms with respect to late presentation and the number of persons with AIDS. The highest rates of late presenters and persons with AIDS (59.8% and 18.7%, respectively) were observed in the PI arm, in comparison to 41.4% and 9.8% in the INSTI arm and 30.3% and 1.0% in the NNRTI arm. In the NNRTI arm, patients had a significantly higher CD4 T cell count at baseline and were less frequently MSM than those in the reference arm (INSTI).
      Table 1Patient characteristics at baseline (societal perspective)


      Age/sex
      Societal perspective
      NINSTI (n = 133)NNRTI (n = 96)PI (n = 107)overall
      Age (y) mean ± SD33640.8 ± 11.439.0 ± 9.740.7 ± 12.440.3 ± 11.3
      Male sex, n (%)336125 (94.0%)87 (90.6%)97 (90.7%)309 (92%)
      HIV-specific variables
      Late presentation, n (%)33655 (41.4%)29 (30.2%)
      Statistically significant difference compared with the INSTI-arm, P < .05.
      64 (59.8%)
      Statistically significant difference compared with the INSTI-arm, P < .05.
      148 (44.0%)
      CD4 cell count (cells/μl) mean ± SD336371.2 ± 206.5459.9 ± 195.3
      Statistically significant difference compared with the INSTI-arm, P < .05.
      331.3 ± 250.0383.8 ± 223.6
      AIDS, n (%)33613 (9.8%)1 (1.0%)
      Statistically significant difference compared with the INSTI-arm, P < .05.
      20 (18.7%)
      Statistically significant difference compared with the INSTI-arm, P < .05.
      34 (10.1%)
      Mode of transmission (%)336
       MSM83.5%75.0%
      Statistically significant difference compared with the INSTI-arm, P < .05.
      77.6%79.2%
       Heterosexual contact9.0%11.5%13.1%11.0%
       IDU/hemophilia/blood transfusion----
       From endemic area-2.1%3.7%
      Statistically significant difference compared with the INSTI-arm, P < .05.
      1.8%
       Unknown6.0%10.4%5.6%7.1%
       Other1.5%1.0%-0.9%
      AIDS indicates acquired immunodeficiency syndrome; HIV, human immunodeficiency virus; IDU, injection drug use; INSTI, integrase strand transfer inhibitor; MSM, men who have sex with men; NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; SD, standard deviation.
      Statistically significant difference compared with the INSTI-arm, P < .05.
      In total, 44 patients did not have a statutory health insurance (including those with private health insurance), leaving 292 patients from 23 centers for evaluation from the statutory health insurance perspective. Prevalence of AIDS-defining illnesses (10.3 %) and mode of transmission (78.4% MSM; 11.3% heterosexual contact; 0% intravenous drug use; 0% hemophilia/blood transfusion; 1.7% from endemic area; 7.5% unknown; 1.0% other) were comparable for both perspectives. Regarding the societal perspective, “unknown” as mode of transmission differed significantly between the INSTI and the NNRTI study arms. In addition, a significant difference was seen in “MSM” and “heterosexual contact” as mode of transmission between the INSTI and the PI study arms. The large proportion of men, and MSM and heterosexual contact being the most and second most predominant route of transmission, reflects the general HIV-infected population in Germany.
      • an der Heiden M.
      • Marcus U.
      • Kollan C.
      • et al.
      Schätzung der Zahl der HIV-Neuinfektionen und der Gesamtzahl von Menschen mit HIV in Deutschland.
      ,
      • Marcus U.
      • an der Heiden M.
      Schätzung der Zahl der HIV-Neuinfektionen und der Gesamtzahl von Menscrhen mit HIV in Deutschland.
      Resource use during the observation period of 24 months, mean numbers, and prices used for calculation of costs from the societal perspective are shown in Table 2. All patients were cared for by HIV-specialized physicians and received ART. Patients (93.5%) also used other non-ART medication(s), 83.0% reported outpatient visits, and 53.6% reported disability owing to comorbidities. Patients in the NNRTI arm did not use inpatient rehabilitation due to the HIV infection, while patients in the PI arm used neither outpatient rehabilitation nor nursing care due to the HIV infection.
      Table 2Resource use during the observation period of 24 months, mean numbers, and prices used for calculation of costs (societal perspective)
      Societal perspective
      N = 336Patients during observation period (%)Mean days
      These averages refer to all individuals who are included in the societal perspective.
      during observation period mean ± SD
      Cost per day in €Source
      INSTI (n = 133)NNRTI (n = 96)PI (n = 107)
      Hospitalization due to HIV4.50.8 ± 4.4-2.2 ± 10.94912
      Costs per contact.
       + 76
      Diagnosis related groups/calculated on the basis of Bock et al
      Bundesministerium für Gesundheit
      Gesetzliche Krankenversicherung. Vorläufige Rechnungsergebnisse 1.-4. Quartal 2016. 2017.
      Hospitalization due to comorbidities20.21.4 ± 5.32.2 ± 8.21.4± 4.9634calculated on the basis of Bock et al
      Bundesministerium für Gesundheit
      Gesetzliche Krankenversicherung. Vorläufige Rechnungsergebnisse 1.-4. Quartal 2016. 2017.
      Outpatient visits due to HIV (excluding HIV-specialized physicians)24.41.0
      Costs per contact.
      ± 2.2
      0.7
      Number of contacts.
      ± 3.1
      0.9
      Costs per contact.
      ± 2.9
      20-83
      Costs per contact.
      calculated on the basis of Bock et al
      Bundesministerium für Gesundheit
      Gesetzliche Krankenversicherung. Vorläufige Rechnungsergebnisse 1.-4. Quartal 2016. 2017.
      Outpatient visits due to comorbidities (excluding HIV-specialized physicians)83.06.7
      Number of contacts.
      ± 8.9
      6.3
      Number of contacts.
      ± 9.4
      5.7
      Number of contacts.
      ± 7.2
      20-83
      Costs per contact.
      calculated on the basis of Bock et al
      Bundesministerium für Gesundheit
      Gesetzliche Krankenversicherung. Vorläufige Rechnungsergebnisse 1.-4. Quartal 2016. 2017.
      Outpatient visits to HIV-specialized physicians100----accounting system information/values from the payment system of German physicians
      Rehabilitation due to HIV (outpatient)0.60.2 ± 1.70.1 ± 1.2-60statutory health insurance/statutory pension insurance statistics
      Rehabilitation due to HIV (inpatient)2.70.5 ± 3.7-1.9 ± 8.2143statutory health insurance/statutory pension insurance statistics
      Rehabilitation due to comorbidities (outpatient)5.11.4 ± 8.82.2 ± 12.50.1 ± 0.860statutory health insurance/statutory pension insurance statistics
      Rehabilitation due to comorbidities (inpatient)2.71.5 ± 14.10.7 ± 5.10.8 ± 4.7143statutory health insurance/statutory pension insurance statistics
      Nursing1.20.1
      Number of months.
      ± 0.6
      0.1
      Number of months.
      ± 0.6
      -440-1774
      Costs per month.
      payment of the statutory nursing care insurance/additional costs for the care-dependent
      Deutscher Bundestag
      Antwort der Bundesregierung auf die Kleine Anfrage der Abgeordneten Elisabeth Scharfenberg, Maria Klein-Schmeink, Kordula Schulz-Asche, weiterer Abgeordneter und der Fraktion BÜNDNIS 90/DIE GRÜNEN–Drucksache 18/13453–Umsetzung der Pflegestärkungsgesetze der Bundesregierung. 2017.
      ,
      • Koopmanschap M.A.
      • Rutten F.F.
      • van Ineveld B.M.
      • et al.
      The friction cost method for measuring indirect costs of disease.
      Antiretroviral medication100---39-49Lauer Taxe
      Other medication93.5---0-12Lauer Taxe
      Disability due to HIV10.712.1 ± 51.71.2 ± 9.412.2 ± 58.8130Federal Office of Statistics
      Disability due to comorbidities53.620.2 ± 46.222.0 ± 42.219.2 ± 39.6130Federal Office of Statistics
      Reduction in/loss of earning capacity due to HIV2.6---130Federal Office of Statistics
      Reduction in/loss of earning capacity due to comorbidities2.0---130Federal Office of Statistics
      HIV indicates human immunodeficiency virus; INSTI, integrase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; SD, standard deviation.
      These averages refer to all individuals who are included in the societal perspective.
      Costs per contact.
      Number of contacts.
      § Number of months.
      Costs per month.
      Table 3 shows the average annual cost per patient from the two perspectives. From both perspectives, expenditures for ART medication represented the principal cost factor. When compared to the reference arm (INSTI), patients in the NNRTI arm accounted for significantly lower costs for hospitalization due to HIV as well as ART from both perspectives. In addition, from the societal perspective, patients in the NNRTI arm generated significantly fewer costs for disability owing to HIV. From both perspectives, patients in the PI arm had significantly higher costs for ART than patients in the INSTI arm. The significant difference in cost for outpatient visits due to HIV between the NNRTI and the INSTI arms from both perspectives was not of economic relevance. From the statutory health insurance perspective, patients in the PI arm caused significantly higher costs for inpatient rehabilitation owing to HIV than patients in the INSTI arm.
      Table 3Calculated annual healthcare costs per patient
      n = 336/292Annual cost in € per patient mean
      These averages refer to all individuals who are included in the perspective.
      Societal perspective
      Proportion of total costs (%)Annual cost in € per patient mean
      These averages refer to all individuals who are included in the perspective.
      Statutory health insurance perspective
      Proportion of total costs (%)
      INSTI (n = 133)NNRTI (n = 96)PI (n = 107)overallINSTI (n = 116)NNRTI (n = 87)PI (n = 89)overall
      Hospitalization due to HIV2690
      Statistically significant difference compared to the INI-arm (P < .05).
      4752581.3%2750
      Statistically significant difference compared to the INI-arm (P < .05).
      4692521.5%
      Hospitalization due to comorbidities4396844475112.5%3376614424662.7%
      Outpatient visits due to HIV (excluding HIV-specialized physicians)2020
      Statistically significant difference compared to the INI-arm (P < .05).
      21200.1%2120
      Statistically significant difference compared to the INI-arm (P < .05).
      22210.1%
      Outpatient visits due to comorbidities (excluding HIV-specialized physicians)1441541231400.7%1261541201330.8%
      Outpatient visits to HIV (specialized physicians)3743903863821.9%3553623533562.1%
      Rehabilitation due to HIV (outpatient)44030.0%00000.0%
      Rehabilitation due to HIV (inpatient)330134560.3%00111
      Statistically significant difference compared to the INI-arm (P < .05).
      340.2%
      Rehabilitation due to comorbidities (outpatient)40643350.2%160060.0%
      Rehabilitation due to comorbidities (inpatient)1095260770.4%4000160.1%
      Nursing39270230.1%-----
      Antiretroviral medication15 98913 888
      Statistically significant difference compared to the INI-arm (P < .05).
      17 461
      Statistically significant difference compared to the INI-arm (P < .05).
      15 85878.8%15 18712 967
      Statistically significant difference compared to the INI-arm (P < .05).
      16 252
      Statistically significant difference compared to the INI-arm (P < .05).
      14 85085.8%
      Other medication8967968188424.2%6385976526303.6%
      Direct costs18 357160 77819 92918 20790.5%16 99514 76218 42316 76596.9%
      Disability due to HIV78677
      Statistically significant difference compared to the INI-arm (P < .05).
      7945862.9%290343602351.4%
      Disability due to comorbidities13131431124713266.6%2583403373071.8%
      Reduction in/loss of earning capacity due to HIV00000.0%-----
      Reduction in/loss of earning capacity due to comorbidities00000.0%-----
      Indirect costs20991507204119119.5%5483746985423.1%
      Total costs20 45617 58521 97020 11817 54215 13619 12117 306
      HIV indicates human immunodeficiency virus; INSTI, integrase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor.
      These averages refer to all individuals who are included in the perspective.
      Statistically significant difference compared to the INI-arm (P < .05).
      The annual healthcare costs from the societal perspective amounted to €20 118 (SD €6451); first year of therapy: €20 911 (SD €8793); second year of therapy: €19 325 (SD €6144). The calculation of costs from the healthcare payers’ perspective yielded €17 306 (SD €4106); first year of therapy: €17 676 (SD €5573); second year of therapy: €16 937 (SD €4045). The highest cost driver due to HIV/AIDS was the cost for ART with €15 858 (SD €1393) from the societal and €14 850 (SD €1306) from the healthcare payers’ perspective. At the time the study was conducted, no generics for the most commonly used antiretroviral drugs were available. Most patients were treated with patented drugs at that time. This accounted for 78.8% and 85.8% of the overall costs, respectively.
      Further costs with significant implications from the societal perspective were as follows: disability secondary to comorbidities (6.6%) €1326 (SD €2793); first year of therapy: €1397 (SD €3375); second year of therapy: €1254 (SD €3173), medications other than ART (4.2 %); €842 (SD €1136) and disability secondary to HIV (2.9 %); €586 (SD €3048); first year of therapy: €832 (SD €4091); second year of therapy: €340 (SD €2352). From the statutory health insurance perspective, these were costs for medication other than ART (3.6%) €630 (SD €899); hospitalization due to comorbidities (2.7%) €466 (SD €1689); first year of therapy: €648 (SD €3051); second year of therapy: €283 (SD €1159); and care from HIV-specialized physicians (2.1%) €356 (SD €107), first year of therapy: €375 (SD €127); second year of therapy: €337 (SD €123). The ART costs in the alternative scenario, taking all changes in regimen into account, were slightly lower with total annual healthcare costs of €20 003 (SD €6290) from the societal perspective and €17 217 (SD €4027) from the healthcare payers’ perspective. Differences in ART costs between the study arms were still statistically significant.
      Statistically significant determinants of annual healthcare costs compared with the reference patient are listed in Table 4. From the societal perspective, the annual cost for the reference patient was €18 251. Cost was higher for patients with AIDS-defining illnesses at baseline (+€7087; 95% CI, €3070-11 589) and were lower for patients with NNRTI as the third agent (−€2183; 95% CI, −€3607 to −€880).
      Table 4Determinants of healthcare costs
      Reference categoriesDeterminants of annual health related costsDifferences in annual health related costs in comparison to the reference case (€)95% CI
      Societal perspective
      No AIDS at baselineAIDS at baseline7087.413070.2411589.18
      INSTI as the third agentNNRTI as the third agent−2183.40−3607.46−880.33
      N1000 samples
      R2/adjusted R20.20/0.17
      Statutory health insurance perspective
      Age (per year)49.9413.3290.34
      No AIDS at baselineAIDS at baseline4493.841726.587587.52
      INSTI as the third agentNNRTI as the third agent−1898.84−2721.86−1020.63
      INSTI as the third agentPI as the third agent1064.3021.062068.30
      N999 samples
      R2/adjusted R20.30/0.28
      Values in Euros. The table shows the significant explanatory variables (P < .05).
      AIDS indicates acquired immunodeficiency syndrome; CI, confidence interval; INSTI, integrase inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor.
      The average cost for the reference patient from the statutory health insurance perspective was €15 148. Higher costs in this model were associated with an AIDS-defining illness at baseline (+€4494; 95% CI, €1727 to 7588) and use of PI as the third agent (+€1064; 95% CI, €21 to 2068). The total costs increased with every year of life at baseline (+€50; 95% CI, €13 to 90). NNRTI as the third agent was associated with lower total costs (−€1899; 95% CI, −€2722 to −1021). Further determinants that were included in the models but did not show any statistically significant influence were female sex, late presentation, and the mode of transmission. The regression explains 20% of the variance of the total costs from the societal perspective and 30% of the variance from the statutory health insurance perspective. Based on the measured VIF, we can assume that there is no multicollinearity in either model.

      Discussion and Conclusions

      The published literature lacks current data on costs associated with HIV infection in Germany.
      • Mostardt S.
      • Hanhoff N.
      • Wasem J.
      • et al.
      Cost of HIV and determinants of health care costs in HIV-positive patients in Germany: results of the DAGNÄ K3A Study.
      ,
      • Treskova M.
      • Kuhlmann A.
      • Bogner J.
      • et al.
      Analysis of contemporary HIV/AIDS health care costs in Germany: driving factors and distribution across antiretroviral therapy lines.
      ,
      • Trapero-Bertran M.
      • Oliva-Moreno J.
      Economic impact of HIV/AIDS: a systematic review in five European countries.
      As previously reported by Mostardt et al (K3A study, reference year 2008) and Treskova et al (CORSAR study, years of data collection 2009-2012), the main cost driver of the healthcare costs were antiretroviral drug expenses. However, the cost of ART found in the PROPHET study was lower than that found by earlier studies.
      • Mostardt S.
      • Hanhoff N.
      • Wasem J.
      • et al.
      Cost of HIV and determinants of health care costs in HIV-positive patients in Germany: results of the DAGNÄ K3A Study.
      ,
      • Treskova M.
      • Kuhlmann A.
      • Bogner J.
      • et al.
      Analysis of contemporary HIV/AIDS health care costs in Germany: driving factors and distribution across antiretroviral therapy lines.
      Aside from considerably falling market prices of ART, this may also be due to different treatment strategies. Patients were switched from more expensive multitablet PI-based regimens to cost-saving single-tablet INSTI and NNRTI regimens during the course of the study. At reduced costs, disease control of HIV remained effective. Subsequently, control of HIV infection and avoidance of secondary diseases achieved with current ART also saved costs in other categories. This finding is supported by the observed decrease in cost of illness in the second year of therapy. Although recent reductions in ART prices may lead to even lower costs of illnesses than the cost calculation with 2017 drug prices, ART remains the most significant cost component of HIV-associated expenditures.
      There were also lower costs due to reduction in or loss of earning capacity as well as HIV-associated hospitalizations compared with Mostardt et al. The reason may lie in differences in methodology. Whereas Mostardt et al included all HIV-infected individuals (ART-naive, first line, second line, third line), the current study evaluated only those initiating ART. In PROPHET, patients were in average about 4 years younger and had considerably less advanced disease. This also reflects the trend of earlier treatment initiation as recommended by the German Austrian HIV treatment guidelines.
      Deutsche AIDS-Gesellschaft
      Deutsch-Österreichische Leitlinien zur antiretroviralen Therapie der HIV-1-Infektion. 2017.
      The method of computing the costs for hospitalization due to HIV was also differed: Mostardt et al used standard resource values, while PROPHET used specific German Diagnosis Related Groups values.
      • Mostardt S.
      • Hanhoff N.
      • Wasem J.
      • et al.
      Cost of HIV and determinants of health care costs in HIV-positive patients in Germany: results of the DAGNÄ K3A Study.
      Compared with Treskova et al, PROPHET found lower costs for hospitalization and higher outpatient care costs, which could be explained by differences in study population and design. Despite a high percentage of late presenters, most patients in PROPHET initiated antiretroviral treatment in an outpatient setting of mostly specialized HIV centers. Furthermore, the pricing methods differed. We confirm the finding of Treskova et al that concomitant medications have a significant impact on costs. Outpatient and inpatient rehabilitation accounted for only minor costs in both studies. Excluding medication costs, Treskova et al do not state whether the reported costs are restricted to those associated with the HIV infection or include total costs.
      • Treskova M.
      • Kuhlmann A.
      • Bogner J.
      • et al.
      Analysis of contemporary HIV/AIDS health care costs in Germany: driving factors and distribution across antiretroviral therapy lines.
      The results of PROPHET underline the findings of Treskova et al and Stoll et al that NNRTI-based ART is linked with cost savings when compared with ART with PI as a third agent.
      • Treskova M.
      • Kuhlmann A.
      • Bogner J.
      • et al.
      Analysis of contemporary HIV/AIDS health care costs in Germany: driving factors and distribution across antiretroviral therapy lines.
      ,
      • Stoll M.
      • Kollan C.
      • Bergmann F.
      • et al.
      Calculation of direct antiretroviral treatment costs and potential cost savings by using generics in the German HIV ClinSurv cohort.
      This is partly explained by differences in pricing of PI-based regimens compared with NNRTI-containing ART. However, the drop in PI price at the end of 2018 may likely have obliterated cost differences between treatment groups. In addition, in this nonrandomized study, we observed significant differences in patient characteristics between PI- and NNRTI-treated patients: in general, PI-treated patients had more advanced disease and lower CD4 cell counts at treatment initiation, subsequently leading to higher costs. Of note, this is also attributed to the fact that rilpivirine, the most commonly used NNRTI, is not recommended for patients with viral loads >100 000 HIV RNA copies/mL. Patients in the NNRTI arm showed significantly lower costs of hospitalization due to HIV and disability due to HIV (only from the societal perspective). This can be explained by significant lower burden of disease in this group (significantly less late presentation, AIDS or severe immune deficiency). This observation highlights the economic importance of early treatment initiation and good disease control.
      A detailed comparison with reported international cost of illness studies is not feasible owing to noncomparable healthcare systems. The systematic review by Trapero-Bertran and Oliva-Moreno points out major differences in the cost of HIV infection in European countries.
      • Trapero-Bertran M.
      • Oliva-Moreno J.
      Economic impact of HIV/AIDS: a systematic review in five European countries.
      In general, international findings agree on rising costs with disease progression and on antiretroviral medication being the main cost category.
      • Trapero-Bertran M.
      • Oliva-Moreno J.
      Economic impact of HIV/AIDS: a systematic review in five European countries.
      ,
      • Boubouchairopoulou N.
      • Athanasakis K.
      • Chini M.
      • et al.
      Cost estimation of HIV infection in Greece: data from an infectious diseases unit.
      • Colombie V.
      • Pugliese-Wehrlen S.
      • Deuffic-Burban S.
      • et al.
      Mean cost of a first combination antiretroviral therapy in HIV-infected patients in France, and determinants of expensive drugs prescription.
      • Tontodonati M.
      • Cenderello G.
      • Celesia B.M.
      • et al.
      Cost of HAART in Italy: multicentric evaluation and determinants from a large HIV outpatient cohort.
      In addition, Krentz et al also report higher costs associated with increasing age of the HIV-positive population.
      • Krentz H.
      • Gill M.
      Increased costs of HIV care associated with aging in an HIV-infected population.
      Lower costs than those reported by Mostardt et al and Treskova et al may also be the result of continuing clinical progress made in the treatment of HIV.
      Some methodical features of PROPHET have to be mentioned. Patients who did not match the inclusion criteria of the analysis (missing documentation) or were lost to follow-up were significantly more often late presenters (55% vs 44%) and differed in the mode of transmission (less MSM (70% vs 79%), more injection drug use (4% vs 0%) and hemophilia/blood transfusion (1% vs 0%). In addition, their mean CD4 cell count was significantly lower (328.2 vs 383.8 cells/μl). Hence, the reported costs in our study from the societal perspective may underestimate the real costs in Germany.
      Independent from changes in the ART regimen, the treatment strategy at the start of the study defined the study group and the calculated cost of ART medication of each patient in the main scenario. Hence, treatment modifications from more expensive PI-based to less expensive regimens were not taken into consideration and would have led to lower overall costs. As shown in the alternative scenario, ART costs in the main scenario are slightly overestimated.
      Another limitation of our study is the fact that the distribution of treatment strategies in our study is not representative of distribution of current first-line antiretroviral regimens in Germany. Patients were recruited along the study arms depending on the ART strategy of their regimen with the aim of achieving 160 patients per study arm. Owing to changes in the availability of new INSTIs and changes in HIV treatment guidelines in 2014, the speed of recruitment in the INSTI arm (fully recruited by December 2014) outperformed the PI and NNRTI arms (recruited by September 2015). Although the use of INSTIs over of PIs is cost-saving, its use over INSTIs instead of NNRTIs is more expensive. Since recruitment into the PI and NNRTI arms was similar, we may assume that both effects balance each other.
      A unique feature of PROPHET is that all patients included were recruited at the time of ART initiation, which allows an analysis of healthcare costs at the beginning up to the second year of therapy.
      To summarize, PROPHET reached the goal of providing a detailed and current overview of the economic impact of HIV based on a large prospective national cohort. It could be shown that ART drug costs and stage of disease determine total costs, which is in line with the findings of other studies. Expenditures on outpatient care are significantly lower than expenditures on ART. However, a reduction in healthcare costs due to ART can be expected with lower prices of newly patented antiretroviral drugs as a result of changing pricing regulations in Germany and with the increasing use of generic antiretrovirals.

      Acknowledgments

      We thank Janssen-Cilag for providing the financial support for the conduct of study. We thank Paul Lauscher from MUC Research, Munich, for data management and data sets preparation. We thank Birgit Mueck from MUC Research, Munich, for monitoring and statistical analyses. We thank all participating patients as well as the staff and investigators of the PROPHET study group: Practice Kreuzberg, H. Schulbin, Berlin; Center for Infectiology Berlin Prenzlauer Berg GmbH (Zibp), A. Baumgarten, C. Mayr, Berlin; MVZ Aerzteforum Seestrasse, W. Schmidt, Berlin; Department of Medicine I, Bonn University Hospital, J. Rockstroh, Bonn; Private Practice Ebertplatz, C. Wyen, E. Voigt, T. Kuemmerle, Cologne; Practice Hohenstaufenring, S. Scholten, S. Schneeweiss, Cologne; Department I of internal medicine, University Hospital of Cologne, G. Faetkenheuer, Cologne; Center for HIV and Hepatogastroenterology, S. Mauss, Duesseldorf; University HIV/STD Center, Essen, Department of Dermatology and Venerology, Universitiy Hospital Essen, S. Esser, Essen; Infektiologikum Frankfurt, M. Bickel, T. Lutz, S. Usadel, Frankfurt; Practice Dr. Ackermann, Halle; ifi-Studien und Projekte GmbH, an der Asklepios Klinik St. Georg, Haus L, A. Stoehr, A. Plettenberg, Hamburg; ICH Studycenter, H.-J. Stellbrink, M. Sabranski, Hamburg; Hannover Medical School, M. Stoll, Hannover; Center for Medical Studies, H. Heiken, Hannover; MVZ Karlsplatz, HIV Research and Clinical Care Centre, Munich, H. Jaeger, Munich; Practice Isartor, R. Pauli, W. Becker, Munich; University Hospital of Munich, J. Bogner, Munich; Department of Medicine II, University Hospital Klinikum rechts der Isar, Munich, C. Spinner, Munich; prinzmed, Practice for infectious diseases, N. Postel, Munich; Center for Interdisciplinary Medicine, S. Christensen, Muenster; Practice Schwabstrasse 26, M. Mueller, A. Ulmer, Stuttgart.

      References

        • an der Heiden M.
        • Marcus U.
        • Kollan C.
        • et al.
        Schätzung der Zahl der HIV-Neuinfektionen und der Gesamtzahl von Menschen mit HIV in Deutschland.
        Epid Bull. 2019; 46: 483-492
        • Mostardt S.
        • Hanhoff N.
        • Wasem J.
        • et al.
        Cost of HIV and determinants of health care costs in HIV-positive patients in Germany: results of the DAGNÄ K3A Study.
        Eur J Health Econ. 2013; 14: 799-808
        • Treskova M.
        • Kuhlmann A.
        • Bogner J.
        • et al.
        Analysis of contemporary HIV/AIDS health care costs in Germany: driving factors and distribution across antiretroviral therapy lines.
        Medicine (Baltimore). 2016; 95e3961
        • US Department of Health and Human Services
        Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in Adults and Adolescents Living with HIV.
        • Rodger A.J.
        • Cambiano V.
        • Bruun T.
        • et al.
        Sexual activity without condoms and risk of HIV transmission in serodifferent couples when the HIV-positive partner is using suppressive antiretroviral therapy.
        JAMA. 2016; 316: 171-181
        • Cohen M.S.
        • Chen Y.Q.
        • McCauley M.
        • et al.
        Antiretroviral therapy for the prevention of HIV-1 transmission.
        N Engl J Med. 2016; 375: 830-839
        • Stoll M.
        • Kollan C.
        • Bergmann F.
        • et al.
        Calculation of direct antiretroviral treatment costs and potential cost savings by using generics in the German HIV ClinSurv cohort.
        PLoS One. 2011; 6e23946
        • Bickel M.
        • Hoffmann C.
        • Wolf E.
        • et al.
        High effectiveness of recommended first-line antiretroviral therapies in Germany: a nationwide, prospective cohort study.
        Infection. 2020; 3: 453-461
        • AOK Bundesverband
        Übersicht über die für 2017 gültigen Landesbasisfallwerte in den einzelnen Bundesländern. 2017.
        • Institut für das Entgeltsystem im Krankenhaus
        G-DRG-Report-Browser 2017.
        Institut für das Entgeltsystem im Krankenhaus, Siegburg, Germany2016
        • Bock J.-O.
        • Brettschneider C.
        • Seidl H.
        • et al.
        Ermittlung standardisierter Bewertungssätze aus gesellschaftlicher Perspektive für die gesundheitsökonomische Evaluation.
        Gesundheitswesen. 2015; 77: 53-61
        • Bundesministerium für Gesundheit
        Gesetzliche Krankenversicherung. Vorläufige Rechnungsergebnisse 1.-4. Quartal 2016. 2017.
        2017
        • Bundesministerium für Gesundheit
        Gesetzliche Krankenversicherung. Endgültige Rechnungsergebnisse 2016. 2017.
        2017
        • Bundesministerium für Gesundheit
        Ergebnisse der Statistik KG 5, Vorsorge- und Rehabilitationsmaßnahmen 2016 der Gesetzlichen Krankenversicherungen. 2017.
        2017
        • Deutsche Rentenversicherung Bund
        Reha-Bericht 2018. 2018.
        2018
        • Bundesministerium für Gesundheit
        Abschlussbericht zur Studie ‘Wirkungen des Pflege-Weiterentwicklungsgesetzes:’ Bericht zu den Repräsentativerhebungen im Auftrag des Bundesministeriums für Gesundheit. 2011.
        2011
        • Deutscher Bundestag
        Antwort der Bundesregierung auf die Kleine Anfrage der Abgeordneten Elisabeth Scharfenberg, Maria Klein-Schmeink, Kordula Schulz-Asche, weiterer Abgeordneter und der Fraktion BÜNDNIS 90/DIE GRÜNEN–Drucksache 18/13453–Umsetzung der Pflegestärkungsgesetze der Bundesregierung. 2017.
        2017
        • Koopmanschap M.A.
        • Rutten F.F.
        • van Ineveld B.M.
        • et al.
        The friction cost method for measuring indirect costs of disease.
        J Health Econ. 1995; 14: 171-189
        • Backhaus K.
        • Erichson B.
        • Plinke W.
        • et al.
        Multivariate Analysemethoden. Eine anwendungsorientierte Einführung.
        11th ed. Springer, Berlin2006
        • Greene W.H.
        Econometric analysis.
        4th ed. Prentice Hall, Upper Saddle River, NJ2000
        • Marcus U.
        • an der Heiden M.
        Schätzung der Zahl der HIV-Neuinfektionen und der Gesamtzahl von Menscrhen mit HIV in Deutschland.
        Epid Bull. 2016; 45: 497-509
        • Trapero-Bertran M.
        • Oliva-Moreno J.
        Economic impact of HIV/AIDS: a systematic review in five European countries.
        Health Econ Rev. 2014; 4: 1-16
        • Deutsche AIDS-Gesellschaft
        Deutsch-Österreichische Leitlinien zur antiretroviralen Therapie der HIV-1-Infektion. 2017.
        2017
        • Boubouchairopoulou N.
        • Athanasakis K.
        • Chini M.
        • et al.
        Cost estimation of HIV infection in Greece: data from an infectious diseases unit.
        Value Health. 2014; 17: A671
        • Colombie V.
        • Pugliese-Wehrlen S.
        • Deuffic-Burban S.
        • et al.
        Mean cost of a first combination antiretroviral therapy in HIV-infected patients in France, and determinants of expensive drugs prescription.
        Int J STD AIDS. 2012; 23: 865-869
        • Tontodonati M.
        • Cenderello G.
        • Celesia B.M.
        • et al.
        Cost of HAART in Italy: multicentric evaluation and determinants from a large HIV outpatient cohort.
        Clinicoecon Outcomes Res. 2014; 7: 27-35
        • Krentz H.
        • Gill M.
        Increased costs of HIV care associated with aging in an HIV-infected population.
        HIV Med. 2015; 16: 38-47