Post-market Review of Chronic Obstructive Pulmonary Disease Medicines ToR 5 Final Report August 2017
PBS/RPBS utilisation analysis methodology
Yüklə 0,73 Mb.
|
- Bu səhifədəki naviqasiya:
- 5.4.2 Department of Health (DoH) PBS claims database
- Drug treatment inclusion and exclusions
- COPD patient cohort
- Estimation of COPD incident and prevalent patients
- Summary steps completed within the analysis
- Data file
5.4 PBS/RPBS utilisation analysis methodology5.4.1 Department of Human Services (DHS) PBS claiming systemA preliminary analysis was conducted to report the total prescriptions and benefits of COPD/asthma medicines subsidised through the DHS PBS claims dataset. Refer to Appendix C for COPD/asthma medicine PBS items included in the analysis. A number of ICS/LABA medications have a Restricted Benefit PBS listing and therefore it is not possible to distinguish between the COPD and asthma populations. All items included in the analysis are above the general co-payment threshold and therefore there was no requirement to add under co-payment source data. 5.4.2 Department of Health (DoH) PBS claims databaseA detailed analysis of the DoH PBS claims data set was conducted based on prescriptions dispensed and unique patients. In order to address some of the issues discussed in Section 5.2 and restrict the data analysis to a population with a higher likelihood of COPD, a number of steps were taken to prepare the DoH PBS data set. Drug treatment inclusion and exclusionsPatients that were supplied their first prescription with an ICS/LABA medicine in the data set period (1 November 2008 to 31 October 2016) were excluded from the analysis. The rationale for the exclusion is to reduce the likelihood of including patients with asthma and mixed airways disease in the COPD cohort and increase confidence that the study has better captured prescribing patterns relevant to COPD patients. A complete analysis of prescribing for COPD would include short-acting reliever medications (SAMAs and SABAs), but analysis of these medications is problematic for several reasons. Firstly, SABAs can be readily purchased over-the-counter in Australia. The extent of utilisation of these medicines is therefore likely to be significantly underestimated on the basis of PBS claims. Secondly, short-acting bronchodilators (and corticosteroids) are inexpensive and priced under the general co-payment. Therefore, these agents were not all captured in the PBS/RPBS dataset prior to April 2012. As all LAMA, LABA, LAMA/LAMA and ICS/LABA preparations are dispensed at a price that is higher than the general PBS co-payment, the PBS/RPBS dataset is complete for the long-acting medicines that are the focus of this Review. Refer to Appendix C for the PBS/RPBS items included in this analysis.
Table 5.1 Standard prescription coverage days calculated from the DoH claims database
Source: PBS Dispensing records between 1/11/2006 and 31/10/2016 file R2017068_PBS_COPD_DATA.CSV. Abbreviations: PBS, Pharmaceutical Benefits Scheme The SCD is used to work out how many days of supply of a medicine a patient has on hand through accumulation and depletion by filling scripts more or less frequently. Breaks in therapy were assumed when a patient does not refill their prescription within two SCD periods after estimated coverage ends. Stockpile adjustments were accounted for in the analysis by estimating the amount of supply on hand, refer to Appendix R for further details. In this analysis COPD initiations were defined as a patient, aged 35 years and older, commencing on medications listed on the PBS for COPD only (LABA, LAMA or LABA/LAMA) that refilled a prescription for the same PBS item at least once within two SCDs. The rationale for the requirement to have refilled is reduce the likelihood of including patients that have received a single supply in the study period and potentially discontinued use due to incorrect diagnosis. A regimen for each one week period was taken to include all medications that were inside their calculated SCDs in that week. Further adjustments were applied to calculate regimens to remove small data artefacts that were not clinically meaningful. Refer to Appendix R for further details regarding the additional rules applied to calculate treatment regimens. The estimated number of COPD patient initiations including regimens are presented in calendar years. Prevalent patients were calculated based on unique patient counts per year for patient regimens that have coverage of any proportion of that year. Patients that switched drug treatments during a year and had more than one regimen were counted as one unique patient. The estimated number of prevalent patients included initiations to COPD therapies and continuing COPD therapy. Patient death was not recorded in the DoH PBS data set and therefore no adjustments were made to the COPD cohort analysis. Patients were assumed to have discontinued treatment if they had not refilled their prescription within 2 SCDs and they were able to re-enter the prevalent pool in a subsequent year if they refilled their prescription at a later date. Patients re-entering the prevalent population were not recounted as initiations. The estimated numbers of COPD prevalent patient are presented in calendar years. No censoring adjustments were made in the final months of the study period to allow for the timing effects of initiation and regimen coverage rules, with the exception of extrapolation, where relevant, to improve comparisons between calendar years (which is noted where it has been used). The method of extrapolation of the 2016 patient analysis was conducted by applying 2015 month on month ratios by class to the corresponding 2016 months for September through December. The rationale for this is to improve transparency of patient behaviour in the later periods that cover the introduction of the more recently PBS listed COPD therapies. An analysis was conducted to determine initiating and prevalent patients on COPD medications including regimens within and outside COPD-X guidelines. A summary of the methodological approach is presented below, for further details refer to Appendix R. Summary steps completed within the analysisStep 1 was to remove the PBS-listed items that are not used in the management of COPD. Step 2 was to exclude patients initiating to ICS/LABA medicine. Step 3 was to retain only COPD patients starting a COPD medicine for the first time during the study period. This is defined as those patients who had not been supplied with medicines used in the management of COPD prior to 1 November 2008 (i.e. applying a look-back period of two years from the first observed PBS/RPBS COPD claim). This is to improve the detection and analysis of initiations (event analysis) and analysis of utilisation (prevalent analysis). Step 4 was to perform a time-to-refill analysis to provide values for use as assumptions for calculating ‘standard prescription coverage days’ (SCDs) for each medicine item. The median time for refill was used in the calculation of the SCD of all COPD medications. Step 5 was to undertake a patient level analysis to determine the estimated medication coverage days for each drug or drug class. This mainly involves detecting breaks in treatment (based upon time-to-refill analysis at Step 4). Step 6 was to calculate patient incidence and prevalence for the selected COPD patient cohort including drug regimens. Step 7 involved estimating the proportion of patients supplied COPD medications within and outside COPD-X guidelines. Data fileSource: PBS Dispensing records between 1/11/2006 and 31/10/2016 file R2017068_PBS_COPD_DATA.CSV refilled at least once (within two SCDs). The SCD values are presented in Table 5.1. Example of analysisFigure 5.1 illustrates the method specified above. Step 5 calculates the SCD for each drug – fluticasone propionate/salmeterol 250/25 μg (FLU/SAL 250/25), tiotropium 18 μg (TIO 18), indacaterol 150 μg (IND 150) and indacaterol/glycopyrronium 110/50 μg (IND/GLY) – arriving at the treatment episodes (blue bars). The Step 6 process (including adjustments) determines the treatment regimens (red bars). In Figure 5.1, the first breaks can be observed in the latter half of 2005. Here fluticasone propionate/salmeterol (FLU/SAL 250/25 – Epi 1) is not refilled on at least two expected supply dates, resulting in a clear therapy break, before being re-continued in 2005 (third bar down FLU/SAL 250/25 – Epi 2). Figure 5.1 also shows (in the second and third blue bars) that tiotropium (TIO 18) and fluticasone propionate/salmeterol (FLU/SAL 250/25) are co-prescribed (and likely co-administered) continuously between late 2005 and late 2013. As a consequence, the treatment regimen is estimated (represented by the red bars) for [FLU_SAL 250/25 with TIO 18] Epi-3. The same logic follows for the remaining individual therapy episodes and the corresponding treatment regimens. Similar methods have been used for assessing medicine use in Australian populations (Pratt et al, 2011; Vitry et al, 2010). Hallas (2005) describes the method and provides references to early variants. This methodology includes a number of the same refinements as implemented by DUSC when analysing concomitant use of medicines in 12-month post-market reviews. Further explanation of the methods is described in Appendix R.
Sample taken from estimated treatment regimens: PBS Dispensing records between 1/4/2005 and 31/12/2015 file REQ198_COPD.CSV Yüklə 0,73 Mb. Dostları ilə paylaş:
|
