Chapter 6. Cost-effectiveness of antimicrobial resistance control policies
The impact of antimicrobial resistance (AMR) on health and health systems expenditure is substantial and set to increase dramatically if no action is taken to curb current trends. Existing AMR control policies have the potential to significantly influence the burden of AMR through first reducing the risk of transmission of infections or by reducing the inappropriate prescription and use of antimicrobials. This chapter reports the findings of a cost-effectiveness model developed to assess and compare the health and economic impact of a number of AMR control policies relative to a business-as-usual scenario in which there are no interventions. The OECD SPHeP-AMR model was used to assess performances of six selected policies – stewardship programmes, improved hand hygiene, enhanced environmental hygiene, rapid diagnostic tests, delayed prescriptions and mass media campaigns – if they were scaled up to national levels in 33 countries. The effects of each AMR control policy on health outcomes and health care expenditure for the 33 countries included in the microsimulation are presented, along with the possible impact of combining different policies. Finally, the strengths and weaknesses of the findings and sensitivity analysis of the main outcomes are discussed.
Note by Turkey:
The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Turkey recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey shall preserve its position concerning the “Cyprus issue”.
Note by all the European Union Member States of the OECD and the European Union:
The Republic of Cyprus is recognised by all members of the United Nations with the exception of Turkey. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.
Key findings
-
The majority of antimicrobial resistance (AMR) deaths can be prevented. Upscaling stewardship programmes, improved hand hygiene and enhanced environmental hygiene policies to national levels would reduce the annual AMR mortality by on average 54% to 58%.
-
Hospital-based AMR control policies can result in up to 2 000 disability-adjusted life years (DALYs) gained per 100 000 persons per year across the included countries
-
The health gains associated with the six control policies tested in the model appear almost immediately following their implementation.
-
Control policies would substantially reduce the number of hospital days due to AMR – up to 800-1 000 hospital days avoided per 100 000 persons, per year for countries with the highest incidence of AMR.
-
For most AMP control policies, implementation costs are largely offset by the savings generated, including for relatively expensive strategies, such as enhanced environmental hygiene, stewardship programmes and rapid diagnostic tests.
-
Improved hand hygiene represents a particularly good investment as its average annual implementation across the countries considered is around USD PPP 8 500 per 100 000 persons for a net return of around USD PPP 140 000.
-
Across all countries, the AMR control policies included in the model deliver high value for money, and for a number of them, represent “best buys” as they generate savings for the health care system.
-
Combining multiple AMR control policies in a broader policy package would generate overall effects (in terms of disease burden and healthcare expenditure) close to the sum of the effects of the individual component policies.
-
A policy package including all three hospital-based policies would save on average USD PPP 1.2 million per 100 000 persons, per year.
-
A community-based policy package and a mixed policy package, would result, respectively, in average reductions in health care expenditure of approximately USD PPP 275 000 and USD PPP 920 000, per year.
6.1. Determining the opportunity cost due AMR
The potential impact of AMR on population morbidity and mortality, health care expenditure and the broader economy is a major cause of concern for health agencies, governments and a variety of policy makers and stakeholders worldwide. The number of recent initiatives and analyses that have attempted to assess the long-term effects of resistance illustrates an increased level of awareness, at the global economic and institutional level, of the threats of AMR. The O’Neil review and the World Bank report on drug-resistant infections, discussed in Chapter 4, are the most recent examples of a series of high-profile government-commissioned publications investigating the potential health and economic burden of AMR. This type of analytical work – which includes the OECD model presented in Chapter 4 – is primarily descriptive. As such, it is fundamental as it provides baseline information that can guide AMR policy and on which further analyses can be developed. In economic terms, however, such evaluations can be considered as partial because they do not provide a complete picture of the potential impact of AMR. More specifically, they fail to estimate the “opportunity cost” associated with AMR – which represents what society is missing out on by committing resources (treatment cost, productivity, life years, etc.) to dealing with the consequences of AMR and not allocating resources to something else. That opportunity cost can only be estimated through the comparison of alternative courses of action or events in terms of both their costs and consequences.
This chapter reports the findings of a cost-effectiveness model (see Box 6.1) developed to assess and compare the health and economic impact of a number of AMR control policies relative to a business-as-usual scenario in which there are no interventions. The objective of the model is to evaluate if the implementation of the selected policies, scaled up to national levels, can reduce the health and economic burden of AMR, the extent of that reduction, if any, and whether those policies would represent efficient investments for governments. The effects of each AMR control policy on health outcomes and health care expenditure for the 33 countries included in the microsimulation are presented, along with the possible impact of combining different policies. Finally, the strengths and weaknesses of the findings and sensitivity analysis of the main outcomes are discussed.
The cost-effectiveness analysis is based on the microsimulation and model parameters described in Chapter 4. The model relies on three main components:
-
A demographic module, which allows the model to account for population demographic evolution (births and deaths). This approach enables the model to reproduce the population dynamics of the included countries.
-
A risk factor module, which includes the exposures of interest. These are the incidence rates of the antibiotic-bacterium combinations included in the study, and the variations of those rates for specific interventions.
-
An outcome module, which produces results from the two first modules and their interactions.
The model described in Chapter 4 constitutes the business-as-usual scenario and its estimates and projections of country-specific outcomes (i.e. AMR mortality, DALYs, health expenditure, and hospital days) were used as the base-case values to which the effects of the selected AMR control policies were compared. The analysis was performed under the assumption that no policies are implemented beyond the policies already in place, and that any potential effects of existing national AMR policies – in terms of reduction in the number of resistant infections – are fully reflected in the AMR incidence data implemented in the model. Figure 6.1 summarises the general approach of the model.
The cost-effectiveness of six AMR control policies and three packages of policies was determined, with specific models for each one of the 33 countries considered, based on the incremental cost-effectiveness ratio (ICER) associated with each policy or package of policies. Outcomes were calculated over the microsimulation period (2015-2050) and a 3% discount rate was applied to future estimates. For all countries, each policy scenario was run 20 000 times in Monte Carlo analyses, to determine 95% uncertainty intervals (UIs) and probabilities of cost-effectiveness.
The cost-effectiveness status of the policies was determined in two steps:
-
1. The value of their ICERs was calculated – as the difference in costs between an intervention and the base-case scenario, divided by the difference in effects between an intervention and the base case. The ratio obtained represents the cost per unit of health outcome (i.e. DALY) (Weinstein and Stason, 1977[1]);
-
2. The probability of cost-effectiveness of each policy was then calculated against the standard threshold of USD 50 000 per health outcome threshold used to determine the cost-effectiveness of health interventions in industrialised economies – i.e. policy makers’ willingness to pay.
Based on these probabilities, the modelled AMR control policies were categorised as either:
-
Cost-saving (leading to an increase in health and net cost-savings),
-
Cost-effective (ICER ≤ USD 50 000 per DALY averted),
-
Non cost-effective (ICER > USD 50 000 per DALY averted),
-
Inferior (i.e. achieving less health benefits and being more costly that the business-as-usual scenario).
This approach (i.e. presenting the probability of cost-effectiveness, instead of the usual display of the ICER associated with each strategy) was preferred as it allows a more informative evaluation and description of the performance of the policies modelled – particularly those with a cost-saving effect.
Country-specific resource needs and implementation costs of the policies were calculated according to the WHO-CHOICE approach. Costs are divided into two main components:
-
Costs incurred at the point of delivery level (most often corresponding to the patient level)
-
Costs incurred at the programme level (corresponding to central costs).
Patient-level costs involve face-to-face delivery by a provider (broadly defined but often corresponding to a health provider) to a recipient - e.g. medicines, outpatient visits, inpatient stays and individual health education messages. Programme-level costs include all resources required to establish and maintain an intervention - administration, publicity, training, delivery of supplies. Interventions like delivery of health education messages on mass media largely involve the latter, while treatment at health centres largely involves the former. A standardised ingredients approach is used, requiring information on the quantities of physical inputs needed and their unit cost (i.e. total costs are quantities of inputs multiplied by their unit costs) (Johns, Baltussen and Hutubessy, 2003[2]) (Johns, Adam and Evans, 2006[3]).
For patient-level costs, quantities are taken from a variety of sources. Where effectiveness estimates were available from published studies, the resources necessary to ensure the observed level of effectiveness are identified. In other cases, the resources implied by the activities outlined in WHO treatment practice guidelines were estimated (WHO, 2003[4]).
Unit costs for each input were derived from an extensive search of published and unpublished literature and databases along with consultation with costing experts.
6.2. How to tackle AMR
As discussed in Chapter 5, the multiple reports of increasing rates of AMR have pushed many stakeholders (governments, health agencies, individual hospitals etc.) to design and implement policies to control the progression of drug-resistance or prevent its occurrence in the first place. These policies can be broadly categorised into two groups:
-
1. Education interventions promoting effective use of antimicrobials – excessive and/or unnecessary use of antimicrobials is a key driver of AMR. The rationalisation of antimicrobial prescriptions would prevent and reduce the emergence of new resistant strains.
-
2. Interventions to prevent the spread of infections – by improving hygiene and reducing the contaminating potential of infected patients.
A variety of AMR control interventions and policies falling in one of these two categories have been described in the literature (Cecchini, Langer and Slawomirski, 2015[5]). However, not all of the exiting approaches to tackle AMR are currently suitable for inclusion in a quantitative modelling framework. The analysis presented here focuses on a subset of policy options for which a strong body of evidence of effectiveness is available. The selection of policies is the result of extended consultations with OECD member countries and relevant stakeholders, primarily members of the OECD Expert Group on the Economics of Public Health. The policies included in the cost-effectiveness model are presented in Table 6.1.
Estimates of the effectiveness of the included policies at the population level (i.e. the scaling up of the intervention to the national level) were identified in the literature and modelled along three main dimensions. The first dimension is the effect of the intervention at the micro-level. For example, a campaign to improve hand-washing practices in hospitals produced a 48% decrease in AMR rates (Kirkland et al., 2012[6]). The second dimension is the potential coverage of the intervention, once it is scaled up. Even if an intervention is implemented nationally, it is likely that some hospitals will not apply it. A common assumption is that between 50% and 70% of health care structures will implement a policy, depending on the income of the country. The third dimension is the time to steady state (i.e. time for the intervention to produce an effect at the population level). Available evidence shows that interventions to tackle AMR require less than one year to reach the steady state (Cecchini and Lee, 2017[7]).
6.3. Healthcare-based interventions
6.3.1. Improved hand hygiene
Insufficient hand hygiene is unanimously recognised as the most important modifiable cause of hospital-acquired infections. Currently, hand-hygiene compliance rates are well below 50% in many OECD countries (Pittet et al., 2004[8]). The implementation of multimodal strategies has been shown very effective in increasing adherence to hand-washing guidelines (Kirkland et al., 2012[6]). The WHO-5 campaign is an example of a strategy promoting hand washing through system change, training and education, observation and feedback, reminders in the hospital, and a hospital safety climate (Kilpatrick, Allegranzi and Pittet, 2011[9]).
Few studies have investigated the effect of hand hygiene on AMR rates. Stone and colleagues calculated a 48% decline in methicillin-resistant Staphylococcus aureus (MRSA) bacteraemia rates following the implementation of the Clean-your-hands campaign (Stone et al., 2012[10]). Similarly, Grayson et al. (2008[11]) and Johnson et al. (2005[12]) concluded that the number of MRSA isolates would decrease by 53% and 40%, respectively. Positive effects were also demonstrated on resistant infections from Escherichia coli (E. coli) and Klebsiella pneumoniae (K. pneumoniae).
The effectiveness of this intervention was modelled by averaging the effectiveness reported by the three studies mentioned above (i.e. -48% of AMR for hospital-acquired infections). The intervention is assumed to be implemented by 70% of the health care structure in a country, affecting all health care-acquired infections (independently of whether they are susceptible or resistant) and its effectiveness is maintained constant over time. The intervention is assumed not to affect the incidence of community-acquired infections.
The intervention is designed as a hand hygiene culture-change programme with the objective to mimic, as much as possible, the WHO-5 campaign. More specifically, the intervention entails some structural changes to ensure that cleaning facilities are available at the point of care (e.g. alcohol-based handrub, soap and water, etc.), combined with training for health care personnel and regular hand-hygiene audits and feedback. Training for personnel and provision of material promoting hygiene is also an important part of the programme.
The cost of the intervention has been estimated at between USD 0.9 and USD 2.5 per capita per year, depending on the country. The calculation of the implementation costs was based on a recent systematic review, which has concluded that the median cost of delivering this type of intervention is around USD 1 per hospital bed per day (Luangasanatip et al., 2015[13]). In addition to the cost at the hospital level, our calculation also includes some costs towards strategic planning at the national level and coordination at the local level (e.g. hospital audit, monitoring, etc.).
6.3.2. Stewardship programmes
Antibiotic stewardship programmes represent a broad category of interventions that usually entail simultaneous implementation of multiple elements, including regulations, guidelines, monitoring, education and campaigns. These actions aim to increase awareness and to rationalise antimicrobial prescription practices among health care personnel. Stewardship programmes are designed and conducted by multidisciplinary teams including specialised physicians (e.g. experts in infectious diseases or microbiologists) and pharmacists.
A large body of evidence supports the effectiveness of stewardship programmes in both hospital and community care settings. A Cochrane review (Davey et al., 2017[14]) showed that implementation of a stewardship programme in a hospital setting decreases both antimicrobial prescription rates (median change up to -40%) and AMR prevalence rates (median change between -24% to -68%, depending on the type of infective bacteria). A more recent systematic review and meta-analysis concluded that stewardship programmes decrease antimicrobial consumption by 19% on average, AMR by 1.7% to 10.4%, and length of hospital stay by 9% (Karanika et al., 2016[15]).
The intervention is modelled on the effectiveness reported by Davey and colleagues (Davey et al., 2017[14]), under the assumption that the average stewardship programme is upscaled to the national level and that uptake from the hospital sector is at around 70% in all the countries included in the analyses. The programme only affects antibiotic-resistant infections acquired in health care settings and its effectiveness is assumed to remain constant over time. Susceptible infections in the hospital setting and susceptible and resistant infections acquired in the community are assumed not to be affected by this intervention.
The implementation of a stewardship programme entails the setting up of a multidisciplinary team including a specialist in infectious diseases, a microbiologist and a pharmacist. The team would review prescription patterns, provide consultation to the other doctors on which antibiotics to prescribe and develop guidelines for antibiotic prophylaxis and treatment aligned with national and international antibiotic guidelines. The team would also deliver specific courses, for example as part of an ongoing medical education programme, to train hospital personnel. As part of the intervention, brochures would be distributed to medical staff and hospital personnel.
The cost of implementing such a programme has been estimated between USD 2.5 and USD 12 per capita per year, depending on the country. The most expensive cost component is the salary of the stewardship team, which accounts for more than 90% of the total costs across countries. Remaining costs cover planning at the central and local level and material for the training of health care personnel in hospitals. The implementation costs vary quite substantially across countries. Wages of health care personnel as well as the number of health care facilities in the country are the most important drivers of such a difference.
6.3.3. Enhanced environmental hygiene
Hospital environments are a reservoir for drug-resistant organisms as pathogens can survive for up to seven months on inanimate surfaces (Donskey, 2013[16]) (Otter et al., 2013[17]). It is estimated that cross-transmission from the environment causes up to 20% of nosocomial infections (Weber, Anderson and Rutala, 2013[18]). Resistant bacteria contaminate objects when they are shed from the skin scales of colonised or infected individuals. These contaminated environmental surfaces and portable equipment such as stethoscopes or pagers can consequently transmit pathogens to susceptible individuals via direct or indirect contact (Blazejewski et al., 2015[19]).
Risk of AMR transmission from contaminated environments depends on the density of resistant pathogens, cleaning practices, patient comorbidities, and the intensity of medical care. Several studies have shown that the risk of acquiring multi-drug resistant organisms (MDROs) is increased if a room has previously been occupied or used by a person colonised or infected by a resistant bacterium (Drees et al., 2008[20]). Systematic cleaning of equipment and the environment is therefore essential for reducing microbial contamination and the subsequent risk of transmission, in particular after discharging an infected individual. Enhanced environmental hygiene can therefore represent a key transmission prevention strategy. This encompasses the decontamination, disinfection, cleaning and sterilisation of the immediate environment and equipment. It further comprises the disposal of items which may have come into contact with infected individuals.
A systematic review was conducted to identify studies evaluating the effectiveness of enhanced environmental hygiene in reducing the transmission of AMR. Thirteen studies met the inclusion criteria and a meta-analysis was performed to calculate pooled estimates of their results. Based on this analysis, enhanced environmental hygiene interventions, depending on the MDRO considered, were estimated to reduce AMR rates by 26% to 49%.
An intervention enhancing environmental hygiene entails any of the following three actions (Donskey, 2013[16]):
-
Disinfectant substitution: This involves a change from detergent to disinfectant, or to a different disinfectant assumed to have higher effectiveness against certain pathogens. The associated cost is estimated at USD PPP 14.2 per person per year.
-
No-touch cleaning: This involves the use of an automated cleaning device, emitting hydrogen peroxide vapour or ultraviolet radiation, to disinfect rooms after routine cleaning. The associated cost is estimated at USD PPP 65 per person per year.
-
Improving effectiveness of cleaning. This may include: additional cleaning time through the employment of new staff; audit, monitoring and feedback regarding cleaning practices and thoroughness; staff education as well as novel techniques of applying products, such as using disposable wipes or colour-coded cloths. The associated cost is estimated at USD PPP 12.6 per person per year.
The average cost of enhanced environmental hygiene interventions was estimated at USD PPP 30.6 per person per year, based on the components of the different types of interventions falling in one of the three categories above.
6.4. Community-based interventions
6.4.1. Delayed antimicrobial prescription
Delayed antimicrobial prescribing avoids unnecessary consumption of antimicrobials in outpatient and primary care settings. Patients are asked to wait up to three days or for a deterioration in their health status before collecting a medical prescription. This strategy both reduces antimicrobial consumption and educates patients that antimicrobials are not always necessary, especially for self-limiting illnesses. Meanwhile, having a prescription provides a sense of safety for both the patient and the clinician if the illness deteriorates (Spurling et al., 2013[21]).
Several reviews and meta-analyses demonstrate that delayed prescriptions reduce antimicrobial prescription rates. Based on the results of clinical trials spanning several OECD countries, Ranji et al. (2008[22]) found the intervention is most effective if patients have to return to the clinic to obtain a prescription if the symptoms do not resolve (Arroll, Kenealy and Kerse, 2003[23]; Spurling et al., 2013[21]).
The effectiveness estimates reported by Ranji and colleagues were used for the analysis. The authors reported that delayed prescriptions could lead to a 50% decrease in the consumption of antimicrobials in the community. A perfectly elastic relationship between consumption of antimicrobials and AMR rates was assumed in the model. This assumption is likely to be conservative as, for many antimicrobials, the relationship is found to be more than elastic (Kaier, 2011[24]). This intervention is also assumed to be implemented by 70% of drug prescribers, affect all the infections developed in the community and its effectiveness remains constant over time. The intervention is assumed not to affect the incidence of infections acquired in hospital settings or of susceptible infections acquired in the community.
The intervention is divided into two components. A first component entails setting up training programmes for general practitioners. General practitioners are given general information about AMR, are explained how to implement the intervention in their practice – the intervention requires the patient to return to the practice three days later if symptoms worsen – and how to respond to potential questions by the patient. A second component is the provision of brochures and posters for general practices. The informative material targets patients and provides information about AMR and the need to decrease antimicrobial consumption. Planning of the intervention is carried out at the national level but its implementation is managed locally.
The cost of the intervention has been estimated at between USD 0.30 and USD 1.30 per person per year. The majority of the resources are devoted to training, followed by the printing and the provision for the information to be handed out to patients. Our analysis does not take into account any economic impact caused by a decrease in the sales of antimicrobials or the extra time needed for doctors to provide a prescription during the second visit, under the assumption that the prescription would be prepared during the first visit.
6.4.2. Mass media campaigns
Raising public awareness about the dangers associated with inappropriate antimicrobial prescription is commonly used across OECD countries as a way to promote rational prescribing. Campaigns are usually delivered during the winter season through mass media, including TV, radio and billboards as well as new forms of media. In some cases, messages are reinforced at the health care service level (mainly in general practices) by providing brochures or leaflets. In many cases campaigns are designed and managed by national health authorities, but sometime the pharmaceutical industry contributes to their development (Huttner et al., 2010[25]).
Few studies have carried out formal quantitative assessments of the potential effectiveness of mass media campaigns. A meta-analysis by Thoolen et al. (2012[26])concluded that mass media campaigns have a small but statistically significant effect on the general population’s attitudes and knowledge towards inappropriate antimicrobial use. By reviewing studies from Italy, the United Kingdom and the United States, Cecchini and Lee (2017[7]) concluded that mass media campaigns may be responsible for a 4% to 9% decrease in antimicrobial prescriptions. Based on these results, mass media campaign interventions were estimated to decrease antibiotic consumption by 6.5%. This was assumed to result in a 6.5% decrease in AMR rates for community-based infections, based on an assumption of perfect elasticity (Kaier, 2011[24]). The intervention is assumed not to affect the incidence of infections acquired in hospital settings or of susceptible infections acquired in the community.
Mass media campaigns are usually delivered over multiple channels, including national TV, radio and newspapers, billboards in major cities and flyers available in general practices and health care structures. Campaigns are modelled in yearly waves, with the media campaign to be reached during the winter season when infections (bacterial or viral) are more likely. A campaign would typically last for the rest of the year but at lower intensity (about half of the advertising coverage compared to the winter period). The cost of such an intervention ranges between USD 1.3 and USD 2.7 per capita per year, depending on the country. The majority of costs are incurred in buying advertising space on the media, followed by the costs of devising and planning the campaign.
6.4.3. Rapid Diagnostic Tests (C-reactive protein tests)
Traditional diagnostic testing methods used to identify bacteria and determine antibiotic susceptibility often require 48-72 hours, limiting their usefulness in guiding antibiotic prescribing and use. New rapid diagnostic tests (RDTs) can produce results within hours that allow clinicians to distinguish viral infections from bacterial infections, identify specific bacterial infections, and determine whether an antimicrobial treatment should be initiated and which drug should be used, based on the identified organism and its susceptibility profile. RDT technologies vary from standard assays to whole genome sequencing approaches and can be used in both hospital and primary care settings. By substantially reducing the time to results, RDTs have the potential to reduce unnecessary antibiotic use (particularly the use of broad-spectrum antibiotics) and improve patient outcomes. Their use is increasing globally.
To derive effectiveness and cost parameters for the RDT intervention in the microsimulation model, we focused on C-reactive protein tests, given the strong evidence in support of this technology and the fact that it can be used for many types of infections – as it effectively distinguishes bacterial from viral infections.
A recent Cochrane review including 9 trials conducted in emergency and primary care setting, reported that that C-reactive protein tests reduce antibiotic prescribing by approximately 25% (Tonkin-Crine et al., 2017[27]). Another review limited to trials conducted in the primary setting reported that use of C-reactive protein tests was associated with a 22% (95% confidence interval: 8% to 44%) reduction in antibiotic prescription. Based on these estimates, RDT interventions were estimated to reduce AMR by 22% in the microsimulation. The intervention was assumed not to affect the incidence of infections acquired in hospital settings or of susceptible infections acquired in the community.
The cost of C-reactive protein tests has been estimated at EUR 11.27 (95% confidence interval: 1.87-24.41 per patient) per patient (Oppong et al., 2013[28]). A study conducted by The National Institute for Health and Care Excellence (NICE) estimated the cost per patient at GBP 12-15 in the United Kingdom (Chaplin, 2015[29]; Hunter, 2015[30]), including the cost of reagents, equipment depreciation and staff time.
6.5. The impact of interventions on AMR
Each of the policies tested in the model seeks to reduce the burden of AMR by affecting one or more factors influencing AMR infections. Figure 6.2 illustrates these factors as well as the specific areas targeted by interventions tested in the model. Target 1 focuses on the selection pressure for AMR created by use of antimicrobials among humans in the community, Target 2 focuses on the selection pressure created by use of antimicrobials in hospital settings, and Target 3 focuses on the impact on bacterial infection rates created by substandard hygiene and living conditions.
Several of the control policies included in the models seek to reduce resistant infections by focusing on Target 1. These policies include delayed prescribing strategies, mass media campaigns, and use of rapid diagnostic tests. The aim of these strategies is to reduce the selection pressure for resistance development and spread among humans in the community by reducing antibiotic consumption in this sector through various means. Lowered selection pressure will then ideally lead to a smaller proportion of resistant bacteria in the community, less transmission of resistance to other sectors, and less global resistance. Ultimately, this decrease in the proportion of pathogens that are resistant will in turn lower the proportion of all bacterial infections that are resistant to antibiotics leading to fewer resistant infections.
Interventions designed to modify Target 2 are based on the same principles as those designed for Target 1 but with a special focus on the hospital setting. As in the community, reducing antibiotic consumption in hospitals will reduce the selection pressure in these environments leading to fewer resistant bacteria and ultimately fewer resistant infections. Interventions focused on this target include stewardship programmes and mass media campaigns.
Lastly, two of the interventions tested, including enhanced environmental hygiene and improved hand hygiene, focus on Target 3. Unlike interventions focused on Targets 1 and 2, these interventions seek to reduce resistance infections by lowering the overall infection rate of both susceptible and resistant infections. This is achieved by reducing the spread of pathogenic bacteria through improved hygiene measures. By targeting all bacteria, these strategies aim to reduce the number but not necessarily the proportion of infections due to resistant bacteria. Interventions targeting bacterial infections may also indirectly affect Targets 1 and 2 by lowering the need for antibiotics in both the community and hospital settings through reduced infections rates.
6.6. The majority of deaths due to AMR can be prevented
All strategies tested in the model lead to reductions in AMR mortality and their effect was consistent across countries. Figure 6.3 shows the estimates per 100 000 persons, for each country, of the annual AMR mortality reductions associated with the six policies tested in the model. The overall preventive effect of the different policies is consistent across all the countries with the improved hand hygiene intervention being the most effective approach. The scale of effect, however, varies significantly between countries. The estimates of the improved hand hygiene policy, for example, range from ten AMR deaths prevented per 100 000 persons annually in Italy, to less than one AMR death per 100 000 persons prevented in the Netherlands or Norway. The differences in scale and magnitude of reductions reflect the heterogeneous epidemiological situation of the included countries, particularly in terms of baseline resistance proportions of infections and rates of change across countries – as discussed in Chapter 3.
Hospital-based policies have the largest impact on AMR mortality. Upscaling the improved hand hygiene, stewardship programmes, and enhanced environmental hygiene policies to national levels would reduce the total annual number of AMR deaths in the included countries by on average 58% (n=37 836), 54% (n=35 270), and 53% (34 931), respectively. These sharp reductions reflect current estimates from studies that have evaluated the included interventions in the clinical context. They are also consistent with the fact that all three interventions have a preventive effect on both resistant and susceptible infections as their impact is based on a reduction of the risk of transmission of all infections, regardless of their resistance status.
Community-based policies are also effective in preventing AMR deaths, albeit to a lesser extent than hospital-based policies. The average annual reduction in AMR mortality associated with the implementation of mass media campaigns, delayed prescription, and RDT is estimated at around 9% (n=5 872), 16% (n=10 214), and 25%, respectively, across the included countries. The lower preventive effect of community interventions on AMR mortality is due to the fact that their implementation would have an impact on resistant infections only. The three community-based policies do not decrease the total probability of acquiring an infection. Their primary effect is a reduction in antimicrobial consumption, which in turn results in a decrease of the probability that an acquired infection is resistant to antimicrobials. This shift, from resistant to susceptible infections, results in a net decrease in total mortality as the probability of dying from a susceptible infection is lower than that of dying from a resistant infection. On the other hand, two out of the three healthcare-based policies (i.e. improved hand hygiene and enhanced environmental hygiene), have an impact on both resistant and susceptible infections, by reducing risk of transmission, and therefore mortality from both types of infections. The stewardship programmes policy – which also affects the level of resistance through reduction in antimicrobial consumption – has a stronger effect on mortality than the community-based intervention because it reduces the probability of resistance of infections acquired in hospital, which are deadlier than those acquired in the community.
Note: IHH: improved hand hygiene, SP: stewardship programmes, EHH: enhanced environmental hygiene, RDT: rapid diagnostic tests. DP: delayed prescription, MMC: mass media campaigns.
Source: OECD analysis based on the OECD SPHeP-AMR model.
6.7. AMR control can deliver substantial health gains
The impact of the interventions in terms of health gains (i.e. DALYs) is consistent with their preventive effect on AMR mortality. Here too, the three healthcare-based interventions considered in the model have the potential of delivering the largest health gains – on average 1 500 to 2 000 DALYs gained per 100 000 persons per year across the included countries. The community-based interventions would generate smaller but meaningful gains at the population level with an average total across countries of 300 to 800 DALYs gained per 100 000 persons per year.
Figure 6.4 shows the evolution of DALYs gained associated with each policy for individual countries. For most countries, the progression tends to be more rapid in the initial phase for healthcare-based policies, while community-based policies gain momentum after five to ten years but deliver substantially less DALYs over time. Again, important heterogeneity exists between countries in terms of scale of the potential health gains, which reflect the baseline epidemiological differences mentioned above. The trends, however, are consistent across all countries, with the improved hand hygiene policy resulting in the highest number of DALYs gained over-time – followed closely by the stewardship programmes and enhanced environmental hygiene policies.
Importantly, the model’s projections show that the potential health returns of AMR control policies would appear rapidly following their implementation. This means that putting in place any of the policies evaluated in the model would deliver virtually immediate health benefits to the population, as shown for all countries in Figure 6.4. This is in stark contrast with most existing population-based prevention policies – such as cancer screening programmes (Hanley, 2011[31]) and other chronic disease prevention initiatives (Cecchini et al., 2010[32]) – which are usually characterised by several years and sometimes decades (in the case of cancer screening for example) of long lead-time before any substantive public health benefit can be observed.
An important aspect of the potential health gains that might result from the implementation at the national level of AMR control policies is the impact on employment and productivity. In the last few years, several evaluations of the potential impact of AMR on the global economy have been published. A report published in 2017 by the World Bank estimated that by 2050, the effects of AMR on labour supply and productivity in the livestock sector could result in a decline in the annual global gross domestic product (GDP) of 1.1% to 3.8%. The shortfall associated with such a contraction was projected to reach USD 1 trillion to USD 3.8 trillion per year after 2030. Increased mortality and its effects on the population size of countries were identified as the largest driver of the decrease in GDP. A detailed evaluation of the macro-economic effects of AMR and AMR control was beyond the scope of the model presented in this report. However, comparing the World Bank estimates to the potential health gains shown in Figure 6.3 and Figure 6.4 sheds light on the capacity of a large-scale implementation of AMR control policies to substantially mitigate any long-term negative effects of AMR in terms of labour supply, productivity, and the global economy.
Note: Future effects are discounted at a 3% rate.
Source: OECD analysis based on the OECD SPHeP-AMR model.
6.8. Reduced burden on hospital resources
The effect of the modelled AMR control policies in terms of hospital days avoided is consistent with the results reported on health outcomes and health expenditure. The model clearly shows that the control policies, if implemented, would lead to large reductions of the number of hospital days due to AMR. However, large differences exist between countries in terms of scale of the reductions.
As for the previous model outcomes, healthcare-based policies would result in the highest number of hospital days avoided. In Italy, for example, the improved hand hygiene policy would result in around 1 000 hospital days avoided per 100 000 persons, each year. The same policy, if implemented in The Netherlands, would lead to approximately 60 hospital days avoided per 100 000 persons. These reductions would represent, for both countries, around 40% reduction in the number of hospital days due to AMR, each year. The difference in effect between the two countries translates differences in the burden of AMR, but likely also in the number of hospital admissions and average length of stay.
Logically, in absolute terms, countries with the highest burden of disease will benefit the most from the preventive effects of the control policies. However, the results also show that health systems with lower levels of AMR (supposedly as a result of more robust AMR control policies already in place) would also draw substantial benefits, in terms of health care resources, from the implementation of AMR control strategies. This applies to both hospital and community-based policies, as shown in Figure 6.5.
Note: IHH: improved hand hygiene, SP: stewardship programmes, EEH: enhanced environmental hygiene, RDT: rapid diagnostic tests. DP: delayed prescription, MMC: mass media campaigns.
Source: OECD analysis based on the OECD SPHeP-AMR model.
6.9. Lower health expenditure
The estimated costs of implementing the modelled AMR control policies and the potential savings over time in terms of health care expenditure are presented in Figure 6.6 and Figure 6.7. Implementation cost estimates are based on the per capita costs reported scaled up to national level by considering key drivers of expenditure including, among others, population coverage, demographic dynamic and health care system arrangements.
For all countries, enhanced environmental hygiene and stewardship programmes appear as the most expensive strategies to implement, with large variations across countries. Italy and the United States, for example, have an average annual cost for enhanced environmental hygiene policies of USD PPP 220 000-250 000 per 100 000 persons, while the implementation costs for that strategy in Australia, Canada and the most northern European countries are well below USD PPP 40 000 per 100 000 persons. The third most expensive policy is the RDT approach for community infections, which comes at an average annual cost of USD PPP 48 000 per 100 000 persons, ranging from USD PPP 1 000 to USD PPP 136 000 per 100 000 across the included countries. These high price tags are consistent with what the strategies involve in terms of staff training and recruitment as well as acquisition of special devices (e.g. no-touch cleaning devices and diagnostic technology in the cases of the enhanced environmental hygiene and the RDT policies, respectively). The three other policies tested in the model, improved hand hygiene, delayed prescription and mass media campaigns, had an average implementation cost of less than USD PPP 15 000 per 100 000 persons per year.
All the policies are expected to result in null or reduced health care expenditure. This means that the savings in health care expenditure resulting from their implementation would almost entirely and immediately offset the implementation costs associated with the different policies, including the ones involving a high initial level of government investment. The improved hand hygiene policy represents a particularly attractive investment. Its implementation cost is on average 10 times lower than the enhanced environmental hygiene policy and generates savings in health expenditure that represent, depending on the country, on average 15 times the implementation costs.
The community-based policies tested in the model – despite the somewhat higher uncertainly around their estimates and more limited effects in terms of savings over-time relative to the healthcare-based policies – would represent good investment for governments as: i) their implementation cost is gradually off-set over-time; and ii) all three policies would prevent AMR deaths and generate health benefits in the population.
Note: IHH: improved hand hygiene, SP: stewardship programmes, EEH: enhanced environmental hygiene, RDT: rapid diagnostic tests. DP: delayed prescription, MMC: mass media campaigns.
Areas of the bars above the zero line represent the health expenditure associated with each policy. Areas below zero represent reductions in health expenditure.
Source: OECD analysis based on the OECD SPHeP-AMR model.
Note: Future savings are discounted at a 3% rate. The figures show the magnitude and evolution of the net savings associated with each control policy over the model’s timeframe.
Source: OECD analysis based on the OECD SPHeP-AMR model.
6.10. The cost-effectiveness of AMR control
Figure 6.10 summarises the results of the country-specific cost-effectiveness analyses performed for the included AMR control policies. It shows the probability of cost-effectiveness of each strategy based on the value of its ICER (see Box 6.1). Across all countries, the implementation of the policies included in the model can be considered as “best buys” as they appear as very likely to deliver high value for money, and even cost-savings, for the health care system. The results of the OECD model are consistent with the available evidence on the cost-effectiveness of AMR control interventions (see Box 6.3). The improved hand hygiene policy would be, by far, the most efficient approach with over 90% likelihood of being cost saving. Its probability of not being cost-effective or inferior to the business-as-usual scenario is close to zero for all countries. The stewardship programme and the enhanced environmental hygiene strategies appear also as highly efficient policies with a very high probability that their implementation will result in health expenditure savings.
The implementation of community-based interventions, despite their more modest effects on health and cost outcomes described above, would deliver high returns on investment for governments. This particularly applies to the delayed prescription and the RDT policies, which are both highly likely to result in cost-savings in a majority of countries. Finally, the mass media campaigns would also represent a cost-effective investment with a very high probability of its implementation leading to cost-savings in the most populous countries.
Several studies have evaluated the cost-effectiveness of AMR control policies. All of the current estimates, however, stem from analyses focusing on “micro” interventions conducted at individual hospitals or health care units to control a specific resistant bacterium. Among the policies included in the OECD model, the stewardship programmes strategy has been evaluated the most, and studies have consistently demonstrated its cost-effectiveness. A recent report commissioned by the Dutch Ministry of Health to identify good AMR control policy practices showed that antimicrobial stewardship teams in hospitals represent a highly effective strategy to tackle AMR as they can lead to more than a 10% reduction in antimicrobial prescribing and cost-savings of over EUR 40 000 (hospital-wide), per year. Similarly, the use of RDTs to regulate antimicrobial prescribing in primary care was associated with a 40% reduction in antimicrobial prescriptions and cost-savings evaluated at EUR 7 per patient treated (Oberjé, Tanke and Jeurissen, 2016[33]).
The cost-effectiveness of improved hand hygiene programmes, at the hospital level, is well established. The WHO Guidelines on Hand Hygiene in Health Care identified several studies, conducted in a variety of countries, which have concluded that hand hygiene initiatives are highly effective in preventing transmission of health care associated infections and typically cost-saving shortly after their implementation (WHO, 2009[34]; Chen et al., 2011[35]).
The OECD model differs from previous analyses because it is the first to adopt a full health system perspective to assess the effectiveness and cost-effectiveness of AMR control policies. Overall, the findings of the model are largely consistent with the available evidence, highlighting the large potential positive impact of these policies being implemented and coordinated at the national level.
Source: OECD analysis based on the OECD SPHeP-AMR model.
6.11. More effective AMR control through combined policies
AMR is a multi-faceted public health issue which requires policy interventions in a variety of settings. The policies assessed in the cost-effectiveness analysis model address resistant infections occuring in both hospital and community settings. Even though the performance of each individual intervention is reported, the objective of the OECD model is not to identify a single or preferred AMR control policy approach. For countries that intend to effectively tackle AMR, it is important to realise that they cannot rely on a single policy to accomplish their goal. OECD analyses show that – as is often the case with prevention policies – the combined effects of multiple AMR control policies results in larger health and economic benefits.
Combining multiple AMR control policies in a broader policy package would generate overall effects (in terms disease burden and healthcare expenditure) close to the sum of the effects of the individual component policies, for all three policy packages tested in the microsimulation. The end result is broadly additive, despite an assumption of less that additive effect. This likely results from the fact that both hospital and community-based policies influence AMR from different angles, which results in a complimentary effect when they are combined. For example, improved hand hygiene reduces the risk of infection transmision from clinical staff to patients, while improved environmental hygiene reduces risk of infection through contact with contaminated objects or surfaces. Similarly, mass media campaigns create awareness in the population, while RDTs reduce the level of inappropriate use of antimicrobials. The two interventions are complementary as they impact the same cause of selection pressure – i.e. inappropriate consumption of antimicrobials – but through different vectors.
Figure 6.9 presents the number of life years and DALYs potentially gained with three possible combinations of AMR control strategies. These include a “healthcare-based policy package” combining improved hand hygiene, stewardship programmes and enhance environmental hygiene; a “community-based policy package” combining delayed prescrition, mass media campaigns and RDTs; and a broad “mixed policy package” combining stewardship programmes, enhance environmental hygiene, mass media campaigns, and RDTs.
The effect of the three policy packages are consistent with the results reported for individual interventions, in terms of scale of the impact of policies as well as the heterogeneity across countries. The healthcare-based package appears as the most effective in terms of health gains with, on average across the included countries, 118 DALYs gained and 10 life years saved per 100 000 persons, each year. Implementing the mixed package would lead to slightly lower health gains with an annual average across counties of 100 DALYs gained and 8 life years per 100 000 persons. As expected, the community-based package would have the lowest level of health benefits with around 32 DALYs gained and 3 life years per 100 000 persons.
The potential reduction in health care expenditure and savings associated with the three AMR control policy packages can be substantial for countries, as shown in Figure 6.10. The model estimates show that all three policy packages represent “best buys” as the health care expenditure associated with their implementation would be largely offset and savings would be generated. Annually, the healthcare-based package would save on average USD PPP 1.2 million per 100 000 persons. The mixed-policy and community-based policy packages, would result, respectively, in average reductions in health care expenditure of approximately USD PPP 920 000 and USD PPP 275 000 per year.
Source: OECD analysis based on the OECD SPHeP-AMR model.
Note: HBI: healthcare-based interventions, MI: mixed interventions, CBI: community-based interventions.
Areas of the bars above the zero line represent the health expenditure associated with each policy. Areas below zero represent reductions in health expenditure.
Source: OECD analysis based on the OECD SPHeP-AMR model.
6.12. Strengths and limitations of the findings
The OECD model relies on many parameters and large quantities of data, derived from a variety of sources. It also relies on several analytical decisions and assumptions made to address knowledge gaps in the epidemiology of AMR in the available costing data etc. (see Chapter 4). Therefore, as for any health-economic evaluation exercise, variations in any of these parameters will have an impact on the performance estimates of the included control policies.
Probabilistic sensitivity analysis was used to evaluate the uncertainty of the cost-effectiveness of each control policy and to generate cost-effectiveness plans showing the distribution of costs and DALYs for each evaluated AMR control policy for individual countries.
The effects of two key assumptions in the model were further tested in separate deterministic sensitivity analyses, and their effects on the probability of cost-effectiveness of the different control policies evaluated:
-
1. the 100% elastic relationship between consumption of antimicrobials and AMR rates, reduced to 80%
-
2. the 70% implementation/adherence to the control policies modelled, reduced to 50%.
Overall, the results of the sensitivity analyses and the relatively narrow uncertainty intervals around most model estimates support the validity of the findings. The reduction of the elasticity assumption and that of the level of adherence to the control policies had a similar and somewhat limited impact on the cost-effectiveness profiles of the different control policies. Under both scenarios, for most countries, the six strategies conserved high probabilities of being cost-effective. The likelihood of cost-savings was reduced considerably for all strategies, with the exception of improved hand hygiene, which would likely generate savings despite reduction in adherence or intervention effect elasticity (see Annex Figure 6.A.1 and Annex Figure 6.A.2).
The model is likely to substantially underestimate the preventive effect of the community-based interventions. This is due mainly to the limited number of community infections included in the model, which again is a result the limited data and understanding of the epidemiology of AMR in the community setting. In addition, our analysis is based on the conservative assumption that a decrease of AMR rates in community infection does not produce any decrease in AMR rates in healthcare-based infections.
An important point to highlight is the health care sector perspective adopted in the cost-effectiveness model. Such a perspective, by definition, fails to assess the indirect costs and benefits generated outside of the health care sector. The model therefore does not provide an assessment of the potential impact of AMR control on the economy as a whole. As such, it likely underestimates the beneficial effects of the control policies tested. Another major limitation of the model is its exclusive focus on human health. It is well documented that the vast majority of antimicrobials produced are destined for the livestock sector where their usage is inadequately regulated for a large number of countries. This results in selection pressure, which could lead to high levels of AMR and risk of transmission of resistant pathogens to humans, which in turn is likely to substantially affect the effectiveness and cost-effectiveness of AMR control policies.
Finally, it should be noted that the objective of the OECD cost-effectiveness model was to provide country specific data on the resources required to implement a set of AMR control policies; to generate information on feasibility, validity, reliability, and cost–effectiveness. The results reported in this chapter should be interpreted as the potential AMR reduction that could be achieved in optimal circumstances. The difficulty of implementation is recognised but beyond the scope of this modelling work
Conclusion: Tackling AMR is an excellent investment
The OECD SPHeP-AMR model is the first attempt to assess the potential impact on health outcomes, health care expenditure and cost-effectiveness of AMR control policies from the perspective of 33 different national health systems. It provides compelling evidence that the widespread implementation of the modelled policies would dramatically reduce the number of deaths due to resistant infections and generate substantial savings for health care systems. It also demonstrates that AMR is not a fatality and that it can be tackled effectively with a variety of strategies that are complementary and readily available.
References
[23] Arroll, B., T. Kenealy and N. Kerse (2003), “Do delayed prescriptions reduce antibiotic use in respiratory tract infections? A systematic review”, British Journal of General Practice, Vol. 53/496, pp. 871-877, https://www.ncbi.nlm.nih.gov/pubmed/14702908.
[19] Blazejewski, C. et al. (2015), “Efficiency of hydrogen peroxide in improving disinfection of ICU rooms”, Critical Care, Vol. 19, p. 30, https://doi.org/10.1186/s13054-015-0752-9.
[5] Cecchini, M., J. Langer and L. Slawomirski (2015), Antimicrobial Resistance in G7 Countries and Beyond: Economic Issues, Policies and Options for Action, OECD, Paris, http://www.oecd.org/els/health-systems/Antimicrobial-Resistance-in-G7-Countries-and-Beyond.pdf.
[7] Cecchini, M. and S. Lee (2017), “Low-value health care with high stakes: Promoting the rational use of antimicrobials”, in Tackling Wasteful Spending on Health, OECD Publishing, Paris, https://doi.org/10.1787/9789264266414-6-en.
[32] Cecchini, M. et al. (2010), “Tackling of unhealthy diets, physical inactivity, and obesity: health effects and cost-effectiveness”, Lancet, Vol. 376, https://doi.org/10.1016/S0140-6736(10)61514-0.
[29] Chaplin, S. (2015), “CRP testing could reduce antibiotic prescribing”, Prescriber, Vol. 26/18, pp. 29-30, https://doi.org/10.1002/psb.1387.
[35] Chen, Y. et al. (2011), “Effectiveness and Limitations of Hand Hygiene Promotion on Decreasing Healthcare–Associated Infections”, PLoS ONE, Vol. 6/11, p. e27163, https://doi.org/10.1371/journal.pone.0027163.
[14] Davey, P. et al. (2017), “Interventions to improve antibiotic prescribing practices for hospital inpatients”, in Cochrane database of systematic reviews, John Wiley & Sons, Ltd, https://doi.org/10.1002/14651858.cd003543.pub4.
[16] Donskey, C. (2013), “Does improving surface cleaning and disinfection reduce health care-associated infections?”, American Journal of Infection Control, Vol. 41/5 Suppl, pp. S12-9, https://doi.org/10.1016/j.ajic.2012.12.010.
[20] Drees, M. et al. (2008), “Antibiotic exposure and room contamination among patients colonized with vancomycin-resistant enterococci”, Infection Control & Hospital Epidemiology, Vol. 29/8, pp. 709-715, https://doi.org/10.1086/589582.
[11] Grayson, M. et al. (2008), “Significant reductions in methicillin-resistant Staphylococcus aureus bacteraemia and clinical isolates associated with a multisite, hand hygiene culture-change program and subsequent successful statewide roll-out”, Med J Aust, Vol. 188/11, pp. 633-640, https://www.ncbi.nlm.nih.gov/pubmed/18513171.
[31] Hanley, J. (2011), “Measuring mortality reductions in cancer screening trials”, Epidemiologic reviews, Vol. 33/1, pp. 36-45, https://doi.org/10.1093/epirev/mxq021.
[30] Hunter, R. (2015), “Cost-Effectiveness of Point-of-Care C-Reactive Protein Tests for Respiratory Tract Infection in Primary Care in England”, Advances in Therapy, Vol. 32/1, pp. 69-85, https://doi.org/10.1007/s12325-015-0180-x.
[25] Huttner, B. et al. (2010), “Characteristics and outcomes of public campaigns aimed at improving the use of antibiotics in outpatients in high-income countries”, The Lancet Infectious Diseases, Vol. 10/1, pp. 17-31, https://doi.org/10.1016/S1473-3099(09)70305-6.
[3] Johns, B., T. Adam and D. Evans (2006), “Enhancing the comparability of costing methods: cross-country variability in the prices of non-traded inputs to health programmes”, Cost Effectiveness and Resource Allocation, Vol. 4, p. 8, https://doi.org/10.1186/1478-7547-4-8.
[2] Johns, B., R. Baltussen and R. Hutubessy (2003), “Programme costs in the economic evaluation of health interventions”, Cost Effectiveness and Resource Allocation, Vol. 1/1, p. 1, https://www.ncbi.nlm.nih.gov/pubmed/12773220.
[12] Johnson, P. et al. (2005), “Efficacy of an alcohol/chlorhexidine hand hygiene program in a hospital with high rates of nosocomial methicillin-resistant Staphylococcus aureus (MRSA) infection”, Med J Aust, Vol. 183/10, pp. 509-514, https://www.ncbi.nlm.nih.gov/pubmed/16296963.
[24] Kaier, K. (2011), “ECONOMIC MODELING OF THE PERSISTENCE OF ANTIMICROBIAL RESISTANCE”, Natural Resource Modeling, Vol. 25/2, pp. 388-402, https://doi.org/10.1111/j.1939-7445.2011.00114.x.
[15] Karanika, S. et al. (2016), “Systematic Review and Meta-analysis of Clinical and Economic Outcomes from the Implementation of Hospital-Based Antimicrobial Stewardship Programs”, Antimicrobial Agents and Chemotherapy, Vol. 60/8, pp. 4840-4852, https://doi.org/10.1128/AAC.00825-16.
[9] Kilpatrick, C., B. Allegranzi and D. Pittet (2011), “WHO First Global Patient Safety Challenge: Clean Care is Safer Care, Contributing to the training of health-care workers around the globe”, International Journal of Infection Control, Vol. 7/2, https://doi.org/10.3396/ijic.v7i2.011.11.
[6] Kirkland, K. et al. (2012), “Impact of a hospital-wide hand hygiene initiative on healthcare-associated infections: results of an interrupted time series”, BMJ Quality & Safety, Vol. 21/12, pp. 1019-1026, https://doi.org/10.1136/bmjqs-2012-000800.
[13] Luangasanatip, N. et al. (2015), “Comparative efficacy of interventions to promote hand hygiene in hospital: systematic review and network meta-analysis”, BMJ, p. h3728, https://doi.org/10.1136/bmj.h3728.
[33] Oberjé, E., M. Tanke and P. Jeurissen (2016), Cost-Effectiveness of Policies to Limit Antimicrobial Resistance in Dutch Healthcare Organisations, http://celsustalma.nl/images/Publicaties/AMR_finalreport_celsusacademie.pdf.
[28] Oppong, R. et al. (2013), “Cost-effectiveness of point-of-care C-reactive protein testing to inform antibiotic prescribing decisions”, British Journal of General Practice, Vol. 63/612, pp. e465-e471, https://doi.org/10.3399/bjgp13X669185.
[17] Otter, J. et al. (2013), “Evidence that contaminated surfaces contribute to the transmission of hospital pathogens and an overview of strategies to address contaminated surfaces in hospital settings”, American Journal of Infection Control, Vol. 41/5 Suppl, pp. S6-11, https://doi.org/10.1016/j.ajic.2012.12.004.
[8] Pittet, D. et al. (2004), “Hand Hygiene among Physicians: Performance, Beliefs, and Perceptions”, Annals of Internal Medicine, Vol. 141/1, p. 1, https://doi.org/10.7326/0003-4819-141-1-200407060-00008.
[22] Ranji, S. et al. (2008), “Interventions to reduce unnecessary antibiotic prescribing: a systematic review and quantitative analysis”, Medical Care, Vol. 46/8, pp. 847-862, https://doi.org/10.1097/MLR.0b013e318178eabd.
[21] Spurling, G. et al. (2013), “Delayed antibiotics for respiratory infections”, Cochrane database of systematic reviews 4, p. CD004417, https://doi.org/10.1002/14651858.CD004417.pub4.
[10] Stone, S. et al. (2012), “Evaluation of the national Cleanyourhands campaign to reduce Staphylococcus aureus bacteraemia and Clostridium difficile infection in hospitals in England and Wales by improved hand hygiene: four year, prospective, ecological, interrupted time series study”, BMJ, Vol. 344/may03 2, pp. e3005-e3005, https://doi.org/10.1136/bmj.e3005.
[26] Thoolen, B., D. de Ridder and G. van Lensvelt-Mulders (2012), “Patient-oriented interventions to improve antibiotic prescribing practices in respiratory tract infections: a meta-analysis”, Health Psychology Review, Vol. 6/1, pp. 92-112, https://doi.org/10.1080/17437199.2011.552061.
[27] Tonkin-Crine, S. et al. (2017), “Clinician-targeted interventions to influence antibiotic prescribing behaviour for acute respiratory infections in primary care: an overview of systematic reviews”, Cochrane Database of Systematic Reviews 9, https://doi.org/10.1002/14651858.CD012252.pub2.
[18] Weber, D., D. Anderson and W. Rutala (2013), “The role of the surface environment in healthcare-associated infections”, Current Opinion in Infectious Diseases, Vol. 26/4, pp. 338-344, https://doi.org/10.1097/QCO.0b013e3283630f04.
[1] Weinstein, M. and W. Stason (1977), “Foundations of Cost-Effectiveness Analysis for Health and Medical Practices”, New England Journal of Medicine, Vol. 296/13, pp. 716-721, https://doi.org/10.1056/NEJM197703312961304.
[34] WHO (2009), WHO Guidelines on Hand Hygiene in Health Care First Global Patient Safety Challenge Clean Care is Safer Care, http://apps.who.int/iris/bitstream/handle/10665/44102/9789241597906_eng.pdf;jsessionid=10F615F176A54EDDA23F6BB2F77EB29B?sequence=1 (accessed on 17 October 2018).
[4] WHO (2003), Making Choices in Health: WHO Guide to Cost-Effectiveness Analysis, World Health Organization, http://www.who.int/choice/publications/p_2003_generalised_cea.pdf (accessed on 4 October 2018).
Annex Figure 6.A.1 presents the probability of cost-effectiveness of interventions relative to the business as usual scenario under the assumption of 80% elasticity between the effect of the intervention on consumption and its impact on incidence of AMR. In the main analysis, the model operates under the assumption of 100% elasticity.
Annex Figure 6.A.2 presents the probability of cost-effectiveness of interventions relative the business as usual scenario under the assumption of 50% adherence to implemented interventions. In the main analysis that assumption is of 70%.
