Applied Health Economics and Health Policy最新文献

Introduction

Antimicrobial resistance is a global emergency related to overprescribing of antibiotics. Few studies have explored how prescribing behaviours may change as the consequence of changing resistance. Understanding how contextual factors influence antibiotic prescribing will facilitate improved communication strategies to promote appropriate antibiotic prescribing. We aimed to develop and conduct a discrete choice experiment (DCE) to measure how contextual factors influence intended antibiotic prescribing of general practitioner (GP) registrars.

Methods

Factors included as attributes in the DCE were level of antibiotic resistance, requirement for an authority to prescribe, existence of a Practice Incentives Program (PIP) for low prescribing and supervisor support for low prescribing. The survey was administered in an online format for GP registrars undergoing training between 2020 and 2021. Regression analysis using a conditional logit model with interaction effects was used on the basis of the assumptions of independence of irrelevant alternatives, independence of error terms and no preference heterogeneity.

Results

In total, 617 unique respondents answered at least one choice set question. Respondents showed significant preference for avoiding prescribing antibiotics when antibiotic resistance was 25–35% or 40–60% compared with 5–8%. There was also a significant preference for avoiding prescribing when an authority to prescribe was required, or when there was supervisory support of low antibiotic prescribing. In the main effects analysis, respondents were significantly less likely to choose a prescribing option if there was a PIP; however, when interaction effects were included in the regression analysis there was a significant interaction between PIP and resistance rates, but the preference weights for PIP was no longer significant.

Conclusions

Knowledge about community resistance impacts the stated intention of GP registrars to prescribe antibiotics. The use of the DCE may have made it possible to determine factors influencing prescribing that would not be detected using other survey methods. These findings provide guidance for producing, explaining and communicating issues regarding antibiotic prescribing to GP registrars.

Introduction

Genomic medicine has features that make it preference sensitive and amenable to model-based health economic evaluation. Preferences of patients, caregivers, and clinicians related to the uptake and delivery of genomic medicine technologies and services that are not captured in health state utility weights can affect the intervention’s cost-effectiveness and budget impact. However, there is currently no established or agreed-on approach for integrating preference information into economic evaluations. The objective of this study was to explore approaches for incorporating preferences into model-based economic evaluations of genomic medicine and to develop a conceptual framework to consider preferences in health economic models.

Methods

We conducted a critical interpretive synthesis of published literature guided by the following question: how have preferences been incorporated into model-based economic evaluations of genomic medicine interventions? We integrated findings from the literature and expert opinion to develop a conceptual framework of ways in which preferences influence economic value in the context of genomic medicine.

Results

Our synthesis included 14 articles. Revealed and stated preference data were used to estimate choice probabilities and to value outcomes. Our conceptual framework situates preference data in the context of health system, patient, clinician, and family characteristics. Preference data were sourced from clinicians, patients and families impacted by a condition or intervention, and the general public. Evaluations employed various types of models, including discrete event simulation, microsimulation, Markov, and decision tree models.

Conclusion

When evaluating the broad benefits and costs of implementing new interventions, sufficiently accounting for preferences in the form of model inputs and valuation of outcomes in economic evaluations is important to avoid biased implementation decisions. Incorporation of preference data may improve alignment between predicted and real-world uptake and more accurately estimate welfare impacts, and this study provides critical insights to support researchers who seek to incorporate preference information into model-based health economic evaluations.

Aim

To systematically review the evidence on productivity losses due to health problems arising from the COVID-19 pandemic based on evidence from population-level studies.

Methods

Following PRISMA statement, we conducted a systematic review using Medline, Embase, Scopus, Web of Science, EconLit, WHO COVID-19 Research and EuropePMC databases and a grey literature search. We included population-level studies using secondary data and qualitatively assessed eligible studies. For a quantitative cross-study comparison, we calculated losses in 2020 international dollars and as a share of gross domestic product. PROSPERO registration number: CRD42023478059.

Results

Thirty-eight studies were eligible for review, most of which reported losses in high-income countries and the European region. COVID-19 was a focus of 33 studies while 3 studies investigated losses from both long COVID and excess mortality. The Human Capital Approach dominated (30 studies) and no study used the Friction Cost Approach. Most studies (84%) reported on premature mortality losses and a quarter provided estimates of losses due to absenteeism. Of the 33 studies eligible for quantitative comparison, we found that the productivity losses ranged from 0 to 2.1% of gross domestic product; the greatest losses were in the high-income countries and for those aged 40–59 years; and losses among men contributed to around 3/4 of the total burden.

Conclusion

The available evidence on the topic is limited, particularly considering the methodological approaches used. Thus, more research is needed to reach a more comprehensive understanding of economy-level productivity losses resulting from the recent COVID-19 pandemic.

Background

Non-invasive prenatal testing has the potential to be a useful genetic screening tool in Australia. However, concerns have been raised about its cost, commercial provision, the psychological impacts of the screening process, and disparities in access experienced by rural and regional communities.

Aims

The aims of this study are (1) to estimate Australian preferences for features of prenatal screening; (2) to explore potential variations in preferences between metropolitan and rural/regional communities; (3) to estimate the extent to which respondents are willing to trade-off between attributes, using willingness to pay (WTP) and willingness to wait estimates.

Methods

A discrete choice experiment (DCE) was conducted with 12 choice tasks. The DCE recruited participants from metropolitan (n = 160) and rural/regional (n = 168) locations across Australia. Mixed logit and latent class analyses were conducted and WTP and willingness to wait were calculated.

Results

Both metropolitan and rural/regional preferences were significantly impacted by the false-positive rate, false-negative rate, and cost. In addition, rural preferences were significantly impacted by the scope of the conditions covered, the inconclusive rate, and wait times. The number of screening tests and revealing the sex of the foetus were not significant within either group. Willingness to pay estimates ranged from AU$13 to avoid a test with a 1% increase in the false-positive rate to AU$323 to screen for a wide range of conditions.

Conclusions

This study highlights the importance of considering differing preferences between rural and metropolitan populations when delivering prenatal screening. Further, this study provides Australian-specific WTP estimates to be incorporated into economic evaluations.

Background

Health technology assessment (HTA) can be conducted at the national, provincial, or hospital level. Although provincial and hospital-based HTAs often focus on non-pharmaceutical interventions, budget impact analysis (BIA) methods for non-pharmaceutical interventions have received less attention in the literature.

Methods

We reviewed HTAs of non-pharmaceutical interventions published since 2015 by a Canadian provincial HTA agency, evaluating the characteristics and challenges of conducting a BIA.

Results

We summarized the unique characteristics of BIAs for different categories of interventions, including surgery and other procedures, diagnostic or screening tests, therapeutic programs, and digital health technologies. We then discussed specific methodological and practical considerations for conducting a BIA of a surgical or other hospital-based procedure. Critical points for BIA methods include the following: (1) when estimating the size of a target population, healthcare system capacity must be accounted for, and historical volumes may offer more realistic figures than prevalence and incidence rates; (2) factors that affect the uptake of a new intervention include guideline recommendations, labor and infrastructure requirements for implementation, and the target population size; (3) when interpreting a budget impact that shows cost savings, analysts must address where the savings are generated from and whether they can be reallocated. Some of the considerations discussed may also apply to HTAs of pharmaceuticals.

Conclusions

When conducting a BIA of a non-pharmaceutical intervention, addressing these methodological considerations may help in better predicting the financial impact of the new intervention for the public payer and guide appropriate budget allocation for healthcare system planning.

Introduction

Healthcare payers in the USA increasingly cover genetic testing, including exome sequencing (ES), for pediatric indications. Analysis of claims data enables understanding of utilization and costs in real-world settings. The objective of this study was to describe genetic test utilization, diagnostic outcomes, and costs for children who received ES as well as for those who received less comprehensive forms of genetic testing, along with their families.

Patients and Methods

We analyzed linked family claims data for commercially insured members of a large regional health plan. The sample included children younger than 18 years of age who had at least 1 year of continuous plan enrollment and at least one claim for genetic testing from 2016 to 2022, as well as their family members. We compared outcomes for children who ever had a claim for ES (ES cohort) with those for children who had claims for only less comprehensive genetic testing (other genetic testing (OGT) cohort). We evaluated the frequency of ICD-10 codes indicating genetic diagnoses, health care utilization, and out-of-pocket costs in relation to the timing of the index genetic test using t-tests and inverse-probability-of-treatment weighted regression models to control for observable clinical and demographic characteristics associated with type of testing received.

Results

Our sample included 182 children (mean comorbidity index 4.78) in the ES cohort and 1789 children in the OGT cohort (3.63; p < 0.001). ES led to an average of 1.44 (95% confidence interval [CI] 0.67–2.20) more new genetic diagnostic codes after testing than OGT. A larger proportion of the proband’s family members had subsequent genetic testing in the ES cohort (mean 33.3%) than in the OGT cohort (0.5%; p < 0.001), but no differences in the number of new genetic diagnoses in family members were observed. Out-of-pocket costs for genetic testing did not differ between the two cohorts stratified by clinical severity.

Conclusions

In our sample of commercially insured pediatric patients, claims for ES were less frequent and occurred among children with more clinical complexity than those for less comprehensive genetic testing. Children in the ES cohort had a higher number of new genetic diagnoses post-testing than those in the OGT cohort with no significant differences in out-of-pocket cost of testing to families. Our findings suggest that ES is being reimbursed for children who may be difficult to diagnose.

Background

Patients with suspected rare diseases often experience lengthy and uncertain diagnostic pathways. This study aimed to estimate the cost-effectiveness of exome sequencing (ES) in different positions in the diagnostic pathway for patients suspected of having a rare genetic disease.

Methods

Data collected retrospectively from 305 patients suspected of having a rare genetic disease (RGD), who received clinical-grade ES and participated in the Canadian multicentre Care4Rare-SOLVE study, informed a discrete event simulation of the diagnostic pathway. We distinguished between tests that can lead to the diagnosis of a specific RGD (‘indicator tests’) and more routine non-RGD diagnostic tests (‘non-indicator tests’). Five strategies were considered: no-ES, and ES as 1st, 2nd, 3rd, or 4th test (Tier 1, Tier 2, Tier 3, and Tier 4, respectively), where ES was the final test in the diagnostic pathway if included. Outcomes included the diagnostic yield, time-to-diagnosis, time on the diagnostic pathway, and test costs for each strategy. The cost-effectiveness analysis from a Canadian healthcare system perspective was conducted with diagnostic yield as the primary outcome of interest. Probabilistic analyses and expert-defined scenario analyses quantified uncertainty.

Results

Implementing ES increases the diagnostic yield by 16 percentage points from 20% with no-ES to 36%. Exome sequencing, as the first test (Tier 1), resulted in the shortest time to a diagnosis and the lowest testing cost. Mean testing costs per patient were CAD4347 (95% CI 3925, 4788) for no-ES, CAD2458 (95% CI 2406, 2512) for Tier 1, CAD3851 (95% CI 3684, 4021) for Tier 2, CAD5246 (95% CI 4956, 5551) for Tier 3 and CAD6422 (95% CI 5954, 6909) for Tier 4, with Tier 1 having the highest diagnostic yield at the lowest cost. The scenario analyses yielded results consistent with those of the base case.

Conclusions

Implementing ES to diagnose patients suspected of having a RGD can result in a higher diagnostic yield. Although a limitation of our study was that the yield for the non-ES indicator tests was estimated using expert opinion due to a lack of available data, the results underscore the value of ES as a first-line diagnostic test, offering reduced time to diagnosis and lower overall testing costs.

Background

Economic evaluation aims to compare the costs and results of health strategies to inform public decision making. Although sometimes suggested, until now no national evaluation agency has recommended formally incorporating the cost of greenhouse gas (GHG) emissions generated by health interventions into the estimation of healthcare costs.

Objective

The objective of this study was to test and discuss the feasibility of estimating and including the contribution of GHG emissions cost to the total cost of a surgical intervention, with the example of robot-assisted total knee arthroplasty (RTA), using a micro-costing approach.

Methods

The study was conducted in June 2022 at the William Morey Hospital (France). Data regarding all of the resources (labor, medical equipment, consumables), as well as energy consumption, staff commuting and waste treatment were collected and valued from the hospital point of view. Greenhouse gas emissions were valued using a cost-effectiveness approach. Several sensitivity analyses were performed.

Results

The mean cost per patient of an RTA was estimated to be €4755.65, of which €152.64 (3.21 %) would be attributable to GHG emissions. The contribution of GHG emissions in the overall cost of a health intervention was highly dependent on the convention used for the price of carbon.

Conclusion

Despite persistent theoretical and practical challenges, adding the estimation of GHG emission costs in the economic evaluation of health interventions may provide institutional decision makers with information that allows them to allocate the public healthcare resources more efficiently.

Background

Women’s preferences for time allocation reveal how they would like to prioritise market work, family life, and other competing activities. Whilst preferences may not always directly translate to behaviour, they are an important determinant of intention to act.

Objective

We present the first study to apply a discrete choice experiment (DCE) to investigate time allocation preferences among women diagnosed with breast cancer and women without a cancer diagnosis.

Methods

Time attributes were paid work, household work, caregiving, passive leisure and physical leisure. An income attribute was included to estimate the monetary value of time. The study took place in the UK and the DCE was completed by 191 women diagnosed with breast cancer and 347 women without a cancer diagnosis. Responses were analysed using a mixed logit model.

Results

Women diagnosed with breast cancer have stronger positive preferences for daily activities compared to women without a cancer diagnosis. They require less compensation (not significant) for an additional hour of paid work (£5.58), household work (£7.92), and caregiving (£8.53). They are willing to pay more for an additional hour of passive leisure (£1.70, not significant) and physical leisure (£13.66, significant).

Conclusion

The heterogeneous preferences for time allocation among women have policy implications and are significant for welfare analysis.