Talk:Epidemiological modelling

Discussions

Argument moved from Talk:Economic evaluation:

----K3: . In the latest published FinIP data for pneumonia the data seem to show a higher than expected effeciveness for PCV10. We want to suggest that this data should be cosidered in light of a recent study in Sweden where PCV10 and PCV13 are both used in a real life setting in the same healthcare system. From this study, based on the Swedish National Inpatient Registry held by the National Board of Health and Welfare, it was found that PCV13 has significanlty lower pneumonia rates in infants and that PCV13 may save more in-patient cost due to pneumonia in infants compared to PCV10. Ref. Berglund A., Ekelund M., Nyman L., poster from ISPPD-9. --Pfizer Oy, 2014-08-27 13:47:39 - R upload in Economic evaluation (type: truth; paradigms: science: comment)

----1: . We are aware of the potential impact of including other than invasive endpoints in the vaccine evaluation. The currently available data are not sufficient to (e.g.) pneumonia as one of the endpoints. Nevertheless, taking into account all evidence, the pneumococcal conjugate vaccination programme is likely to be cost saving in Finland. The implications of this on vaccine comparison will be discussed elsewhere. Kauq, Sep 7, 2014] (type: truth; paradigms: science: comment)

Questions:

• Vaccine efficacy on non-typeable Haemophilus influenzae cannot be taken into account as comparison criterion because neither of the vaccines have prevention of NTHi in their indication. Moreover, for both vaccines the Summary of Product Characteristics (SmPC) approved by European Medicines Agency states in section 4.4., Special warnings and precautions, that the vaccines do not provide protection against other bacteria: Synflorix: There is insufficient evidence that Synflorix provides protection against pneumococcal serotypes not contained in the vaccine or against non-typeable Haemophilus influenzae. Synflorix does not provide protection against other micro-organism Prevenar 13: Prevenar13 will only protect against Streptococcus pneumonia serotypes included in the vaccine, and will not protect against other microorganism that cause invasive disease, pneumonia or otitis media. Based on the same sentences in SmPCs of the vaccines, the vaccines do not provide protection against other serotypes than those contained in the vaccines. Therefore, adjustability of the serotype composition in the economic assessment (http://en.opasnet.org/w/Economical_assessment ) is purposeless and the user defined option should be removed.

⇤--1: . Possible effects of pneumococcal conjugate vaccines on other than pneumococcal endpoints are not taken into account in comparing pneumococcal vaccines. This is due to lack of sufficient evidence to include such endpoints. The two currently licensed vaccines may induce protection against some serotypes not included in the vaccine because of cross-reactive immunity (see sensitivity analyses on pages Epidemiological modelling and Economic evaluation).

Regarding the question about including other than current vaccine serotypes in the epidemiological modelling tool, we have retained the possibility to adjust for arbitrary serotype compositions to ensure a generic approach to vaccine comparison. Moreover, the model serves as a platform to learn about a simple mechanistic approach to describe trends in serotype replacement, beyond the use of the currently available vaccines. Of note, the modelling tool uses the two currently available vaccines as defaults. In the actual comparative analysis, additional assumptions on direct efficacy against invasive pneumococcal disease for some serotypes are included (see Economic evaluation). Kauq, Sep 6, 2014 (type: truth; paradigms: science: attack)

• "The focus is on the incidence of invasive pneumococcal disease (IPD) cases in different age groups covering the whole population."

⇤--#: . Because the epidemiological model neglects the benefits of prevention of respiratory infection episodes, which effectively represent majority of the vaccine preventable disease burden (e.g. Palmu et al, ESPID, 2014), the outcomes are considered insufficient to inform the cost effectiveness analyses, unless supplemented with estimates for reductions in the respiratory illness, including pneumonia and acute otitis media endpoint. -- GSK 12:33, 3 September 2014 (UTC) (type: truth; paradigms: science: attack)

⇤--1: . The above comment is true. Pneumococcal conjugate vaccination is clearly more cost effective than inferred on the basis of invasive disease alone. In fact, it is likely to be cost saving. However, in the absence of reliable data on the vaccine-preventable incidence of respiratory disease, we have deferred including respiratory infections in the analysis for the time being. See also our responses to the corresponding discussion on page Evaluation_criteria. Kauq, Sep 6, 2014 (type: truth; paradigms: science: attack)

• "The coverage of vaccination and vaccine efficacy against carriage are assumed to be high enough to justify the assumption of complete elimination of vaccine-type carriage among both the vaccinated and also, due to substantial herd effects, among the unvaccinated members of the population."

⇤--#: . The assumption that vaccine type (VT) disease will be reduced to zero, at a steady state, in children, should be extended to VT-related types, where effectiveness has been demonstrated.

For adults and elderly, a recent review by Feikin et al, Plos Medicine, 2013, which has reviewed almost a decade of experience with PCVs in UMV has shown that while there was a dramatic reduction (nearly elimination) of VT/VT-related disease in children, in adults and elderly, the VT continue to cause disease years after implementation of the programme. Therefore, the assumption about complete elimination of VT disease in these age groups, even at a steady state, is not supported by available data.

Assuming the adult disease is reduced to zero would be a gross overestimation of what can be realistically expected from a PCV programme based on the experience with PCVs so far. The assumption should be therefore adjusted based on the findings reported by Feikin et al, Plos Medicine, 2013, except where there is clear evidence suggesting no impact of the vaccine on nasopharyngeal carriage for a particular serotype. --GSK 12:33, 3 September 2014 (UTC) (type: truth; paradigms: science: attack)

⇤--1: . Even assuming the time frame of 5-10 after vaccination onset, it may be true that full elimination is not obtained for some serotypes, even if there would be indirect effects through reduced transmission. Then again, the available data on both PCV10 and PCV13 vaccines have not accumulated for long enough to make any definitive statements.

The current model approach includes some cross-reactive serotypes for the two currently available conjugate vaccines, in particular serotypes 6A and 6C (see the discussion on ‘Evaluation criteria’). A sensitivity analysis was performed to investigate the importance of the assumptions on vaccine comparisons. Kauq, Sep 6, 2014 (type: truth; paradigms: science: attack)

• "Vaccine-type carriage will be completely replaced by carriage of the non-vaccine types whose disease causing potential is not altered by vaccination."

⇤--#: . There is some uncertainty with respect to the assumption of the extent of replacement for respective vaccine formulations (e.g Vesikari, ECCMID, 2013, Dagan, CID, 2013). Because this could extensively impact on the outcomes of the model, this uncertainty should be acknowledged and preferentially also be reflected in the weight which is attributed to the outcomes of the modelling in the award criteria. --GSK 12:33, 3 September 2014 (UTC) (type: truth; paradigms: science: attack)

⇤--1: . There is considerable evidence about the impact of serotype replacement on the effectiveness of PCV vaccination. The same assumptions have been made regarding the 10 common serotypes in the two currently available pneumococcal conjugate vaccines. Regarding the three additional serotypes in PCV13 (3, 6A, and 19A), special attention has been given to evaluate the impact of assumptions regarding their efficacy (direct protection) or expected effectiveness (indirect protection). This includes e.g. possible cross-immunity against 6A for PCV10 as well as possible cross-immunity against 6C for PCV13. Kauq, Sep 6, 2014 (type: truth; paradigms: science: attack)

• "The replacement model was built to reflect the accumulated 15 year long experience on use of pneumococcal conjugate vaccines worldwide and the related scientific research activity."

----#: . Given the limitations and uncertainties around the assumptions of the model, especially on the replacement and consecutively net impact of PCVs in adults and elderly, as highlighted in the previous comments, it should be considered to take a conservative approach and assume similar replacement in disease and consequently similar net effect in adults and elderly with both formulations, that was observed with the first generation PCV, as a base case of the model. --GSK 12:33, 3 September 2014 (UTC) (type: truth; paradigms: science: comment)

⇤--1: . We have not adopted this approach. Some important differences between the two vaccines regard the effect of serotype replacement in the adult population. The current sensitivity analysis indicates that, depending on the assumptions, either of the two currently available vaccines may be more effective in the general population. Kauq, Sep 6, 2014 (type: truth; paradigms: science: attack)