The causal association between cervical cancer and Human Papilloma Virus (HPV) is one that is well established (3)(4)(5). Thus, prevention strategies are largely targeted at preventing HPV infection or preventing disease progression for those who are infected. There are 2 types of preventive measures available to reduce incidence and mortality from cervical cancer: i.) Vaccination and ii.) Screening. While vaccination is a primary preventive measure (providing protection against the incidence of illness), screening is a secondary preventive measure aiming to diagnose illness early and prevent its progression. Combining screening and vaccination against HPV should potentially provide the best protection against cervical cancer as neither option alone offers 100% protection. At present, screening strategies for cervical cancer have not been altered for females who are HPV vaccinated (6). Screening vaccinated women is arguably still a requirement because of the limitations of current HPV vaccines both in their lack of therapeutic effect (not protecting women with ongoing neoplastic processes) and in their coverage of limited number of HPV types (leaving to evolve some 25–30% of cervical cancer cases related to HPV types other than 16 or 18 strains). Consequently, for health economists, the question regarding the most cost-effective combination of screening strategies along with vaccination arises. The economic impact of screening HPV vaccinated populations is analytical information that health policy makers require for the formulation of effective, evidence-based strategies.
The purpose of this literature review is to collect and collate the best possible evidence available to answer this question. This review aims to systematically analyze health economic studies on HPV vaccination to provide integrated evidence and recommendations based on its cost–effectiveness when combined with differing cervical cancer screening strategies.
CAVEAT:
Prior Knowledge:
During the search conducted in July 2017, it was noted that a systematic review by Mendes et al (7) on CEA of prevention strategy combinations against HPV infection, was published on March 28th 2017 (after the preliminary literature review search was conducted by the author). Upon examining this paper, it was found that:
i) No quality appraisal of the papers included in the review was carried out.
ii) No papers analyzing the cost effectiveness of screening strategies in populations vaccinated with the non-avalent (9-valent) vaccine (8) were included
iii) The search was finalized in April 2014 resulting in the exclusion of all papers since 2014 till July 2017
iv) The study focused only on studies based in Austria, Belgium, Switzerland, Czech Republic, Germany, Denmark, Spain, Finland, France, Greece, Ireland, Italy, the Netherlands, Norway, Poland, Portugal, Sweden, Slovenia, and the UK, the US, Canada and Australia excluding relevant studies from other parts of the world
Contribution from this literature review:
i) Complete appraisal of all papers using the recommended CHEERS checklist for economic evaluations (9)
ii) A crucial CEA conducted on screening strategies within cohorts vaccinated with the nonavalent or 9-valent vaccine (10)
iii) Studies published after April 2014 were added to this literature review (8) (10)(11)
iv) Relevant studies carried out in Africa (12), Thailand (13), Laos (14), China (15), Taiwan (16) , Israel (17) have been included and reviewed
HPV and Cervical Cancer:
HPV (Human Papilloma Virus) is currently the most common sexually transmitted virus (3). It is passed on primarily through genital contact (such as vaginal and anal sex) and also by skin-to-skin contact (3). Over 100 types of HPV have been identified and more than 40 of these infect the genital area. Although there are several high-risk HPV types, the infection of 2 particular HPV types: 16 and 18 are found to be responsible for 70%–75% of all cervical cancers and 40%–60% of its precursors (18). Among the cancer-related outcomes of HPV infection, cervical cancer is the most important outcome, with over 5,00,000 new cases and 2,75,000 attributable deaths world-wide in 2008 (19). The high-risk (cancer causing) types of HPV include: 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59 and 68 (20). Women who are infected with HPV may have their infection clear, progress or persist. Persistence is the most significant determinant of whether or not a HPV-infected woman will develop a clinically significant sequelae (4).
Primary Prevention – Vaccines:
Currently 2 first generation HPV vaccines have enabled prevention against the two most common types of human papillomavirus infection – strains 16, 18 (Bivalent) and also 6 and 11 (Quadrivalent). Clinical trials have provided evidence that the bivalent vaccine was 100% [95% confidential interval (CI) 47–100%, N = 1113] effective against HPV types 16 and 18, and the quadrivalent vaccine 98% [95% CI 86–100%, N = 10,565] (21) (22). These vaccines, have been made available in several countries since their international approval and we are now in the phase of possibly of utilising the second generation of ‘non-avalent’ HPV vaccines (23). These newer vaccines have potential for protection against persistent infection from HPV types 6/11/16/18/31/33/45/52/58 (which together cause ~90% of cervical cancers, globally) (5).
Secondary Prevention – Cervical Cancer Screening Tools:
Persistent HPV infection can cause pre-cancerous cervical lesions and invasive cervical cancer thereafter. With regular cervical cancer screening and appropriate follow-up, most cervical cancer precursors can be identified and treated, interrupting progression to the severe disease stage. Screening programs can reduce cancer risk among those who do not receive the vaccine, those infected by non-vaccine targeted HPV types, and those who demonstrate reduced vaccine efficacy, providing insurance at the population level, given the uncertainties surrounding vaccine performance. The screening tools available include (24):
a.) Cervical Cytology:
Evidence suggests that cervical cancer screening using cytology is one of the most successful public health prevention programs, particularly when implemented in a mass strategy (25). This method involves the collection of exfoliated cells from the cervix followed by microscopic examination of the same after staining, for cellular abnormalities. Based on these abnormalities, further investigation is required to arrive at a diagnosis. There are two cytological screening methods:
i) Conventional PAP smear
ii) Liquid Based Cytology (LBC)
These 2 methods have been extensively compared and it is found that LBC is significantly advantageous in terms of sensitivity and possibility for ‘reflex testing’ of HPV infection without new sample requirements. (26)(27)(28)
b.) Visual Inspection:
Visual inspection of the cervix continues to be implemented as a screening tool for low-resource settings, despite its limited specificity and poor po
sitive predictive value (PPV), as it is economical and can provide immediate results. This metho
d involves direct visual examination for abnormalities or lesions in the cervix after staining with acetic acid or Lugol’s iodine referred to as Visual Inspection with Acetic Acid (VIA) or Visual Inspection with Lugol’s Iodine (VILI), respectively. (29)
c.) HPV DNA Testing:
Similar to Cytology, HPV DNA testing is done on sample cells collected from the cervix and is used to confirm the presence of infection by the HPV Virus (3)(25). This test has shown higher sensitivity and reproducibility of results than Cytology, for detecting high-grade cervical intraepithelial neoplasia (CIN), (although lower specificity) (30)(31). Unfortunately, the high cost of this option has limited its availability in Low and Middle income countries of the world (32).
Economic Evaluation:
With the costs of healthcare increasing worldwide, there is advancing pressure to ration and efficiently use limited resources. Economic evaluations are utilized to identify, measure, and compare health care intervention costs and benefits, to aid in efficient resource allocation (33).
Economic evaluation, as per the Drummond et. al definition, refers to “the comparative analysis of alternative courses of action in terms of both their costs and consequences” (34). The basic types of economic evaluation include:
1. Cost-benefit analysis (CBA):
CBA measures the benefits and costs of outcomes achieved from an intervention in purely monetary terms. This means that expected years of life gained or expected improvements in health and wellbeing are valued in terms of currency. There is much controversy surrounding CBAs on account of ethical and practical reliability questions on how one may accurately express health outcomes in monetary terms.
2. Cost-effectiveness analysis (CEA):
To overcome the limitations of CBA, CEA was developed as an alternative, more practical approach to healthcare decision-making. It assesses the net cost of a project or service relative to the outcomes generated. CEA is used where the need for a project has already been established, but uncertainty remains over the best method for achieving it (35). The purpose of cost effectiveness analysis is to identify the best method to spend a set budget to achieve a particular goal (36).
3. Cost-utility analysis (CUA):
CUA is often seen as a special form of CEA that introduces measures of benefits that reflect individuals’ preferences over the health consequences of alternative programs that affect them. CUAs use a global measure of health outcome, such as quality-adjusted life-years (QALYs) by undertaking one program instead of another, and the results are often expressed as a cost per QALY gained. (37) This enables the comparison of different types of programs, which makes CUA more practical for decision-makers.
4. Cost-Minimization Analysis (CMA):
Sometimes a cost-minimization analysis is performed if the alternatives under evaluation are considered to achieve the identical health outcomes and carried out in terms of net cost comparisons (38). This specificity of outcome measure reduces its application across health fields, thereby limiting its utilization in evaluation studies.
Source:https://www.ispor.org/PEguidelines/source/Guidelines_Austria.pdf
Economic Modelling:
Full economic analyses of interventions can be carried out by the following approaches:
1. Trial-based studies:
As randomized clinical trials are a necessary condition for the successful licensing of pharmaceuticals, relevant economic data are often obtained alongside the trials for economic evaluations. This method provides internal validity, while the main limitation is that the results may suffer from external generalizability (39).
2. Decision analytic modeling:
This approach brings together a range of evidence sources and allows the expansion of the comparators considered in the analysis and an expansion of the time horizon beyond that of a trial period. Further, decision analytic modeling provides a framework for informing specific decision-making under conditions of uncertainty by allowing more convenient assessment of modeling assumptions, modeling structural uncertainty, and different patient subgroups (heterogeneity) (37). Important model types include:
i) Decision Trees: This is the simplest form of decision analytical modelling in economic evaluation. The pathways in decision trees follow each intervention or process option in a series of logically ordered alternative events, denoted by branches emanating from chance nodes (circular symbols). The alternatives at each chance node must be mutually exclusive and their probabilities sum exactly to one. The end points of each pathway, denoted by terminal nodes (triangular symbols), are assigned values or pay-offs, such as costs, life years, or quality adjusted life years (QALYs). Once the probabilities and pay-offs have been entered, the decision tree is “averaged out” and “folded back” (or rolled back), allowing the expected values of each option to be calculated.
ii) Markov Model: An alternate form of modelling is the Markov model. Unlike decision trees, which represent sequences of events as a large number of complex pathways, Markov models involve simpler and more flexible sequencing of outcomes, including recurring outcomes, through time. Patients are assumed to reside in one of a finite number of health states at any point in time and make transitions between those health states over a series of discrete time intervals or cycles. The probability of staying in a state or moving to another one in each cycle is determined by a set of defined transition probabilities. The definition and number of health states and the duration of the cycles will be governed by the decision problem (40). The final stage is to assign values to each health state, typically costs and health utilities (41)(40). Most commonly, such models simulate the transition of a hypothetical cohort of individuals through the Markov model over time, allowing the analyst to estimate expected costs and outcomes. This simply involves, for each cycle, summing costs and outcomes across health states, weighted by the proportion of the cohort expected to be in each state, and then summing across cycles (42). If the time horizon of the model is over one year, discounting (34) is usually applied to generate the present values of expected costs and outcomes.
iii) Microsimulation models: These models simulate the progression of individuals rather than hypothetical cohorts. They track the progression of potentially heterogeneous individuals with the accumulating history of each individual determining transitions, costs, and health outcome. Unlike Markov models, they can simulate the time to next event rather than requiring equal length cycles and can also simulate multiple events occurring in parallel.
iv) Discrete event simulations: They describe the progress of individuals through healthcare processes or systems, affecting their characteristics and outcomes over unrestricted time periods. These simulations are not restricted to the use of equal time periods or the Markovian assumption and, unlike patient level simulation models, allow individuals to interact with each other.
v) Dynamic models: These models allow internal feedback loops and time delays that affect the behaviour of the entire health system or population being studied. They are particularly valuable in studies of infectious di
seases, where analysts may need to account for the evolving effects of factors such as herd im
munity on the likelihood of infection over time, and their results can differ substantially from those obtained from static models.
Economic Evaluation Outcomes and Decision rule:
The results of an economic evaluation of an intervention are typically expressed in terms of an ICER – Incremental Cost Effectiveness Ratio. ICERs (measured most often in cost per QALY gained) reflect the incremental cost required to sustain one unit of benefit gained from a particular intervention compared to another. It applies to a decision rule based on a threshold cost effectiveness ratio. This decision rule states that any intervention with a price per unit effectiveness above a fixed threshold, would not be implemented and any program with an ICER below the threshold would be implemented. The threshold that this decision rule is applied to differs between economic settings. The threshold recommended by WHO involves utilizing a value which is a multiple of the GDP of the country under study (43) for cost effectiveness acceptability. An alternative to this is to cite the cost–effectiveness of an intervention that has previously been implemented in the country under study and to utilize the same as a benchmark for acceptable cost–effectiveness. The latter, however, is an approach used mainly in High income countries(43).
AIMS:
The aim of this systematic literature review is to present the comprehensive results of all available international evidence on the cost-effectiveness analysis of different cervical screening strategies for HPV vaccinated populations.
OBJECTIVES:
The main objectives of this review are to:
1) Identify studies conducted to examine cost effectiveness of screening carried out for women post-HPV vaccination.
2) Examine and compare cost effectiveness outcomes of different screening strategies based on frequency, tools implemented and age of primary screening.
3) Conduct a critical appraisal of the literature included for the review.
4) Provide an assessment of the reporting quality of the literature included for the review.
METHODOLOGY:
Ethical Approval:
Ethics approval was first applied for on 31st March 2017 after an initial literature search was conducted to confirm that there were no existing systematic reviews on the topic. Approval was granted by the London School of Hygiene & Tropical Medicine MSc. Research Ethics Committee, on 10th April 2017 (Ethics Ref: 13528 /RR/7584).
PICO Framework:
The P.I.C.O. framework was implemented for the formulation of the appropriate researchable question (44):
Population (P): HPV Vaccinated women
Intervention (I): Screening or vaccination or none
Comparators (C): Comparator screening strategy
Outcome (O): Incremental Cost Effectiveness Ratio
Based on the PICO framework for the research topic, the literature review was then carried out following the PRISMA flowchart in phases of Identification, Screening, Eligibility and Inclusion (45)
Search Strategy:
A comprehensive literature search of peer-reviewed, published journal articles in English was carried out in the standard online databases EMBASE, MEDLINE, PUBMED, NHS EED and Cochrane Library. (NHS EED was not a separate search as it is covered through the Cochrane database (46)). The search strategy was designed using appropriate MeSH and Text words to cover synonyms, combinations and word choices with the main categories which included: 1) Cost effectiveness 2) Screening and 3) HPV Vaccination. This strategy was developed with the help of expert advice from the librarians at the London School of Hygiene & Tropical Medicine. The strategy used for the key words were based on an exploded list of associated MeSH words (identified on PubMed) and free text words, as below:
i. Screening: “Early Diagnosis”, “Early Detection of cancer”, “Screening”,
ii. HPV vaccination: “Papillomavirus Vaccine” “Human Papillomatous Vaccine”, “HPV Vaccine” and “HPV Vaccination”
iii. Cost-effectiveness: “cost effective*” “cost-effective*” “costeffective*” “cost-benefit analysis”, “costbenefit analysis”, “cost benefit analysis”, “cost”, “economic”, “benefit”, “effectiveness”, “Incremental cost-effectiveness analysis”, “Incremental cost-effectiveness ratio”, “ICER”.
Boolean commands of “OR” and “AND” were used appropriately, to join synonyms and string the key words together, respectively (47). The bibliographies of selected publications were scanned and titles cross-referenced to ensure relevant studies were not missed out in the database search.
Study Selection:
All titles of the papers identified were reviewed to filter those which were obviously irrelevant. Following this, the titles and abstracts of remaining papers were reviewed and duplicates were removed. Applying exclusion and inclusion criteria (agreed upon by the author and supervisor) papers with content relevant to the research topic were then isolated. Finally, the citations within these papers were screened thoroughly using the same inclusion criteria to ensure all relevant articles were included for review. The final list of papers identified was then examined in full text, for the data extraction process. The search was completed in July 2017.
Inclusion Criteria:
1. Primary economic evaluations which satisfy the Drummond et al. definition of CEA “the comparative analysis of alternative courses of action in terms of both their costs and consequences”
2. Cost effectiveness analysis (CEA) of different screening strategies combined with HPV vaccination
3. Cost effectiveness analysis with outcome parameters expressed in terms of Incremental Cost Effectiveness Ratio (ICER)
4. Articles in the English language available in full text
Exclusion Criteria:
1. Partial economic analyses (studies that consider either costs or consequences but not both) were not included
2. CEAs of cervical screening between vaccinated and unvaccinated cohorts with no explicit analysis of different screening methods within the vaccinated cohort and only minor variations in the sensitivity analysis were not included.
3. CEAs comparing the same screening strategies combined with different HPV vaccine types, schedules and doses were not included
4. Economic analyses which do not provide outcomes in terms of ICERs were not included
DATA EXTRACTION:
Subsequent to literature identification and screening, the data extraction was conducted by filling in an excel sheet with pre-determined fields which included : Authors’ names, Year of research, Geographical context, Aim, Model implemented, Economic perspective, Vaccine parameters (type, effect duration, dose/ schedule and cost), Screening parameters (tool combinations, frequency and starting age), WTP threshold, Time Horizon, Outcome parameter, Base case results, DSA/ PSA results, Scenario analysis results. Data regarding cost effectiveness of interventions analysing vaccination alone
or screening alone was not extracted as it was irrelevant to the research question for this review.
ANALYSIS:
The data extracted was analysed and a narrative description based on their reporting quality, methods and results, grouping them into categories was undertaken.
Quality appraisal:
A quality analysis on reporting of economic evaluations was carried using the recommended CHEERS (Consolidated Health Economic Evaluating Reporting Standards) checklist (9) consisting of 24 items. This checklist was employed because it provides the most relevant criteria for assessing economic evaluations (9) under the subsections of Title and Abstract, Introduction, Methodology, Results and Discussion. The papers were appraised using the checklist version created in Excel and completed in August 2017. Details of the appraisal conducted are attached in the Appendix (Ref. Table No. 2)
Analytical categories:
Owing to high levels of heterogeneity between papers in terms of screening strategy comparisons, modelling methods chosen and geographical context, a descriptive analysis was undertaken. By studying the data extraction tables, papers were then grouped together based on the following broad categories to highlight differences and similarities within these subgroups:
1) ECONOMIC SETTING
2) METHODOLOGICAL APPROACHES
1. Economic perspectives
2. Economic models implemented
3. Outcome measures
4. Cost Effectiveness Thresholds
3) POLICY FINDINGS
1. Screening tool comparisons
2. Screening frequency comparisons
3. Comparisons of varying age of first screening
4. Screening strategies in the context of the nonavalent vaccine
RESULTS:
CAVEAT:
Prior Knowledge:
During the search conducted in July 2017, it was noted that a systematic review by Gervais et al. (48) on CEA of prevention strategy combinations against HPV infection, was published on March 28th 2017 (after the preliminary literature review search was conducted by the author). Upon examining this paper, it was found that:
i) No quality appraisal of the papers included in the review was carried out.
ii) No papers analysing the cost effectiveness of screening strategies in populations vaccinated with the non-avalent (9-valent) vaccine (8) were included
iii) The search was finalized in April 2014 resulting in the exclusion of all papers since 2014 till July 2017
iv) The study included only on studies based in Austria, Belgium, Switzerland, Czech Republic, Germany, Denmark, Spain, Finland, France, Greece, Ireland, Italy, the Netherlands, Norway, Poland, Portugal, Sweden, Slovenia, and the UK, the US, Canada and Australia excluding relevant studies from other parts of the world
Contribution from this literature review:
i) Complete appraisal of all papers using the recommended CHEERS checklist for economic evaluations (9)
ii) 2 crucial CEAs conducted on screening strategies within cohorts vaccinated with the nonavalent or 9-valent vaccine (10)(49)
iii) Studies published after April 2014 were added to this literature review (8)(10)(11)
iv) Relevant studies carried out in Africa (12), Thailand (13), Laos (14), China (15), Taiwan (16) , Israel (17) have been included and reviewed
SEARCH STRATEGY RESULTS:
A total of 1750 studies were identified using the search strategy described above, all of which were examined and filtered to arrive at a final 21 studies which fulfilled all pre-determined inclusion and exclusion criteria illustrated in the PRISMA flow diagram below. The entire search strategy and end results are listed in Table 1 of the Appendix .
PRISMA FLOW DIAGRAM
Quality Appraisal Results:
All of the papers included in this study satisfied most sections in the CHEERS checklist however none of the papers satisfied all of the reporting criteria. Table 2.1 below provides summarized results of the quality appraisal conducted while a detailed extraction is provided in the Appendix Table no. 2. The results demonstrate that the major areas of under-reporting seem to be heterogeneity of populations analysed and currency conversion explanations. It would be important to note that some studies failed to name the type of economic study (i.e. cost effectiveness analysis) in the title, which is essential for proper indexing(50).
TABLE 2.1: Summarized Quality Appraisal Results
Checklist
Item
n %
(Total n=21)
References
Articles reporting 100% of CHEERS
checklist items
0%
–
Missing details:
Title: Economic evaluation description
24%
(8)(51)(52)(15)(10)
Methods: No explicit mention of perspective
5%
(53)
Results: Missing reporting of heterogeneity
90%
(53)(51)(54)(55)(56)(57)(52)(58)(59)(16)
(60)(61)(13)(17)(8)(62)(63)(12)(64)
Methods: Missing description of data sources
5%
(57)
Methods: Missing description of currency conversion
90%
(10)(53)(51)(54)(55)(56)(57)(52)(58)(59)(16)(15)(17)(8)(62)(60)(61)(14)(49)
Other: Missing source of funding
29%
(63)(53)(52) (59)(16)(17)
Other: Missing conflicts of interest disclosure
9%
(51)(52)
ECONOMIC SETTING:
The World bank list of economies, 2016 was used as a reference in labelling the economic setting of the country(65). Contrary to the previous systematic review conducted which only included High income countries, this review included 17 studies based in High Income Countries (HIC’s) while the remaining 4 were based in Low (14)(12) and Middle (15)(13) Income Countries (LMIC’s).
METHODOLOGICAL APPROACHES:
1) Economic perspectives:
From the review conducted, it was found that only 7 studies (17) (10) (53) (23) (11) (64) (61) adopted the Health Service perspective by including only direct medical costs of vaccination, screening and cervical cancer treatment borne by the paying party. The remaining 14 studies adopted the Societal perspective including both direct and indirect medical costs associated with vaccination, screening and cervical cancer treatment.
2) Decision analytical models implemented:
i.) Markov Model: 8 out of the 21 studies (53)(51)(57)(52)(59)(16)(13)(17) included implemented the Markov Model
ii.) Microsimulation model: 5 studies implemented the Microsimulation model (54)(55)(56)(8)(60) all of which were in high income settings and suggested
HPV DNA testing as a cost-effective tool for cervical screening among HPV vaccinated cohorts with varying age of initiation and intervals for screening.
iii.) Dynamic model: Only 3 studies implemented the Dynamic model (10)(14)(49)
iv.) More than one model: A total of 5 studies implemented more than one analytical model using hybrid combinations of a transmission cum disease history model (61)(58), dynamic cum cohort simulation models (15), dynamic model cum deterministic multi-cohort model (11) and natural history cum cohort simulation models (12).
3) Outcome measures:
i. Cost/QALY: 9 of the studies expressed the ICER in terms of Cost Per Quality Adjusted Life Year (QALY) gained (53)(51)(58)(57)(56)(8)(60)(61)(49)
ii. Cost/YLS: 9 of the studies expressed the ICER in Cost Per Year of Life Saved (YLS) (54)(55)(52)(66)(15)(13)(11)(10)(12)
iii. Cost/DALY averted: 2 studies expressed the ICER as Cost / DALY averted (17)(14)
iv. Cost/ QALE: 1 study alone expressed the ICER as Cost / Quality Adjusted Life Expectancy (59)
The choice to represent the ICER in Cost per QALY gained was carried out in all High Income Country settings while the 2 studies with ICERs in Cost per DALY averted were in one Low income setting (63) and one high income setting (17). 9 studies which used Cost per YLS to represent the ICER, were from HIC (54)(55)(52)(66)(15)(10)(11) and 2 were from LMIC (12)(13) settings.
4) Willingness To Pay or Cost-Effectiveness thresholds:
The threshold values adopted for Cost Effectiveness included the:
i) WHO recommendation(34) of 1 to 3 times GDP (54)(15)(13)(17)(60) (61)(14)(12)
ii) Country specific accepted thresholds (59)(58)(52)(57)(56)(55)(51)(53) (16)(8)(10)
Only one study did not specify the threshold adopted in their analysis (11).
POLICY FINDINGS:
1) Screening Tool comparisons:
Different screening tools and tool combination comparisons were made in 10 of the studies reviewed (54)(55)(56)(67)(66)(8)(11)(60)(63)(12). For example, the study by Kim et al (54) examined the cost effectiveness of Cytology alone, Cytology with HPV DNA triage and co-testing with Cytology & HPV DNA while the study by Lew et al (11) examines Conventional cytology, Manually-read liquid-based cytology, Image-read liquid-based cytology, HPV Testing with liquid-based cytology triaging, HPV testing with partial genotyping for HPV 16/18 and liquid-based cytology triaging, Co-testing with both HPV Testing and Liquid based cytology.
2) Screening frequency comparisons:
Cost effectiveness analysis by comparing different intervals between screenings was carried out in all of the studies included for review. The intervals were varied in terms of number of years of gap between one screening strategy in all of the studies except for 3: in which different intervals were applied to different screening tools (17), different intervals were applied to different age groups (11) and different intervals applied to different vaccines types (60) were examined.
3) Comparisons between age of screening implementation:
Different ages of screening initiation and cessation were evaluated in all studies included, except for 3 (66)(13)(58). 9 studies included assessed cost effectiveness cohorts by varying age of initiation of screening alone after vaccination (61)(63)(53)(51)(54)(55)(56)(57)(52)(8). One study by Coupé et al (60), in contrast, examined 3 scenarios with age variations for vaccines of different valences and doses concluding that 4 rounds of HPV DNA screening between 30 and 60 years of age for bivalent vaccine and one lifetime screening for broad spectrum vaccinations proved cost effective in the Netherlands.
4) Screening strategies in the context of HPV 9 / Nonavalent vaccine:
Two crucial studies included in this review (which were not part of the review by Gervais et al. (48)) examine the cost effectiveness of screening women vaccinated with the new nonavalent HPV vaccine in HIC settings.
i.) The study by Simms et al, August 2016 (10) implemented a well-validated dynamic model of HPV transmission and cervical screening incorporating the influence of vaccination, herd immunity and screening. The authors concluded that if the intensity of screening programs is significantly less per woman’s lifetime (post vaccination), screening will remain cost-effective once the nonavalent vaccine is implemented.
ii.) The study by Simms et al, December 2016 (49), examined the cost effectiveness of screening Nonavalent vaccinated cohorts compared to HPV 4 vaccinated cohorts. The authors concluded that screening of the former group proves cost effective, compared to the latter, provided the additional cost per vaccine dose remains between 23 and 36 AUS$.
5) Analysis of the effect of Herd Immunity:
Notably, most of the studies did not account for heterogeneity of the population examined except for 2 which examined the effects of herd immunity within vaccinated populations. Areas worthy of exploring heterogeneity could include population subgroups of varying socio-economic backgrounds, risk exposure viz.: multiple sexual partners or previous history of sexually transmitted infections.
The entire data extraction table built and developed for the purpose of this review is provided in the Appendix in Table 3, for further reference.
DISCUSSION:
In total, 21 studies (available in full text format) published until July 2017 in peer reviewed journals were included in the review. All of the papers were described in the abstracts as cost-effectiveness analyses. The data gathered from the papers are discussed below following the thematic sequence of the results demonstrated above.
QUALITY APPRAISAL:
The CHEERS checklist is recommended for the appraisal of economic evaluations with the aim of providing set guidelines to authors, editors and reviewers to improve reporting standards (50). The studies included in this review were found to be of high quality fulfilling most of the reporting criteria.
Title Quality:
The title of the paper is required to identify the study as an economic evaluation and preferably also mention the type of economic evaluation conducted. This is primarily for the purpose of ensuring the studies are indexed appropriately (50) and there are less chances of the studies being missed by reviewers. 5 studies included in this study failed to provide an accurate title requiring the author to examine the abstract to confirm if indeed it was an economic evaluation.
Abstract Reporting Quality:
CHEERS recommends the use of ‘structured’ abstracts as studies have proven that these provide higher reporting quality than descriptive abstracts, which allows readers to locate relevant information easier (68)(69). While all studies provided structured abstracts, only one study out of 21 provided a structured summary by mentioning all of the specified headings: objectives, perspective, setting, methods (including study design and inputs), results (including base-case and uncertainty analyses) and conclusions.
Quality of Introduction:
This section is expected to provide an explicit statement and explanation of the broad context of the study and its relevance in health policy. This checklist item was scrutinized in conjunction with checklist items 4 till 7 i.e. target population and subgroups, setting and location, study perspective and comparators, as recommended (50). All of
the studies included in this review were found to satisfy the reporting standards for the introduction.
Qual
ity of Reporting Methods:
i) Target population and subgroups: Definition of population groups under evaluation is of particular importance in economic studies as cost-effectiveness results vary by cohort characteristics (70). All the studies in this review were found to provide a detailed explanation of population characteristics with respect to demographic profiles, vaccine coverage or uptake rates, follow-up visit adherence, etc. and explained reasons for assumptions regarding vaccine efficacy, screening coverage etc.
ii) Setting and Location: All studies clearly described the system within which the vaccine and screening interventions were being provided. This ensures that the evaluation undertaken addressed the question relevant to the population setting.
iii) Study Perspective: All evaluations made clear mention of the economic perspective implemented and provided reasons for the choice and costs included, except for one study by Sopina et al, for which the author derived the perspective by examining the costs included.
iv) Time Horizon: Since only preventive interventions were examined in these analyses, they are particularly sensitive to the time horizon (71) owing to influencing factors such as waning of immunity from vaccination, protection from herd immunity with increased coverage, discounting etc. All the studies provided a description of time horizon choice. It was specially noted that all studies utilizing the dynamic model provided explanations for time horizon chosen, as this model is known to be sensitive to this parameter.
v) Measurement of preference based outcomes and data sources: All studies provided specific mention of the preference-based outcome measurements and other data sources except for one by Goldie et al (57), which did not include clear mention of preference-elicitation techniques (e.g. EuroQoL5D), data sources or methods for extrapolating data from published studies.
vi) Reporting of heterogeneity and currency conversion: It is recommended that authors must report differences in costs, outcomes, or cost-effectiveness that can be explained by variation between subgroup characteristics of patients such as age, socio-economic background, co-morbidities etc. The reasons for heterogeneity were poorly touched upon in a majority of studies here. This could be due to poor quality and/or lack of data available, incompatibility with model type to incorporate heterogeneity influences, etc. The cause for poor reporting of currency conversion, on the other hand, could be due to lack of consistent or set guidelines for reporting these elements.
Missing Generalizability Description:
None of the studies included in this review explicitly accounted for generalizability of their results although the findings, limitations and current knowledge explanations were discussed in detail. Economic evaluations must pay particular attention to this criterion on account of the fact that if they are set with a particular jurisdiction in mind, it would be inappropriate to use the findings in a different setting (E.g., LMIC and HIC settings). Authors should ideally, explicitly discuss how findings can and cannot be applied to local or global settings to avoid errors in evidence-based policy formulation.
Missing source of funding and conflict of interest disclosure:
Some of the studies failed to disclose all sources of funding (funds received both directly and indirectly) which limit credibility. This is because studies funded by pharmaceutical companies, for example, may only be interested in publishing findings that may be in favour of interventions requiring products manufactured by them as against non-commercial funders. It is advised that the ICJME recommendations (72) are followed for declaration of conflicts of interest by the authors even when there are no conflicts to ensure transparency and credibility.
ECONOMIC SETTING:
This review included all papers relevant to the research question irrespective of the geographic setting of the analysis, in contrast to the review by Gervais et. al (48). Comparative analysis of economic evaluations across diverse economic resource settings are known to be challenging but in accordance with the pattern observed by Griffiths et al (73) it was found that studies in Low and Middle income countries relied upon micro-costing method for cost data while those in High income countries utilized Gross costing based on national administrative databases.
METHODOLOGICAL APPROACHES:
Choice of Economic Perspective:
Majority of the studies chose the Societal perspective. None of the LMIC based studies utilized the Health service perspective except for one (63) study based in Lao. This is in agreement with the guidelines used by researchers in LMIC’s (73)(74) which state the preferred perspective would be Societal except in Egypt). Adopting a societal perspective, including all direct and indirect costs and benefits, facilitates policies aimed at maximising the welfare gains to society, or minimising the losses (75). However, in practice, it is not always feasible to include all possible costs and benefits in an economic evaluation which is why the health service perspective is popular among researchers. In the context of this review, the use of different perspectives poses a limitation towards comparability of the studies included.
Choice of Decision Analytical model:
Decision modelling for Cervical Cancer is particularly challenging as it involves incorporating elements of both an infectious disease (HPV infection) as well as a chronic illness (Cervical Carcinoma). As there are no clearly defined guidelines on choice of decision model and limited guidance on good modelling (76) each model type examined in this review is discussed in the context of the illness under study. The majority of studies reviewed implemented the Markov Model, all of which make clear their source of cost, utility and transition probability data. This model is simple to develop and is ideal for chronic disease conditions. However, the studies using this model faced the limitation that interaction between individuals or groups (e.g. herd immunity) cannot be accounted for, different characteristics of individuals / groups cannot be incorporated and resource constraints cannot be considered (77). Some studies implemented the Dynamic model, which requires a great volume of parameters and advanced statistical programming knowledge. It is best suited for strategies aimed at controlling infectious diseases. This allows for studying interactions between individuals / groups and estimation of direct and indirect effects of interventions (e.g. herd immunity) (77)(78)(49). The studies which implemented the Microsimulation Model benefited from being able to incorporate individual characteristics changing over time, resource constraints and allowing for random events (77).