Death, be not proud

The release of Australian mortality data by the AIHW brings to mind Donne’s sonnet, although causes of death are no longer the ‘kings and desperate men‘ attributed in c.1610.

In 2017, there were 160,909 deaths registered in Australia, 66% among people aged 75 or over (60% males and 73% females). The median age at death was 78 years for males and 85 years for females. 

Overall, cancer (neoplasms) was the leading cause of death. At 29%, clearly easing circulatory diseases (27%) out of the top spot for the first time. The leading cause of death for males was coronary heart disease (13%). Dementia and Alzheimer disease was the leading cause of death for females (11%), followed by coronary heart disease (10%). Cerebrovascular disease (which includes stroke), lung cancer and chronic obstructive pulmonary disease (COPD) made up the top 5 leading underlying causes of death in Australia in 2017 for both males and females of all ages combined.

In 2017, the difference in death rates between the sexes was the narrowest ever recorded. The greater than 70 % reduction from a high in 1968, is considered to be due to a drop in deaths from circulatory diseases. Factors at play include improvements in medical care (surgery, diagnosis and pharmaceuticals), and lifestyle changes (smoking, diet and high blood pressure).

Child (aged 0-4 yrs) mortality accounted for < 1 % of all deaths in the period. A major improvement on the 26 % reported for 1907, however still too high for the families involved. Work continues on access to and quality of neonatal health care; community awareness of risk factors; and increasing coverage of universal immunisation programs.

As well as differences by gender, the leading causes of death also vary by age (refer graphic):

  • Among infants, perinatal and congenital conditions accounted for 79 % of deaths;
  • Land transport accidents were the most common cause (11 %) among children aged 1–14 .
  • Suicide was the leading cause of death among people aged 15–24 (35%), followed by land transport accidents (22%);
  • For people aged 25–44, it was also suicide (21%), followed by accidental poisoning (12%);
  • Chronic diseases feature more prominently among people aged 45 and over. Coronary heart disease was the leading cause of death for people aged 45–64, followed by lung cancer;
  • For people aged 65–74, it was lung cancer followed by coronary heart disease;
  • Dementia and Alzheimer disease was the second leading cause of death among people aged 75 and older, behind coronary heart disease.
Leading underlying causes of death, by age group, 2015–2017 (AIHW Figure 3.2)

Image: An Unclean Death by Pia Guerra

Survival by cancer stage at diagnosis

Cancer Australia continues to develop the National Cancer Control Indicators (NCCI) resource with a new data release presenting population-level national data on relative survival by stage at diagnosis for the 5-year period 2011-2016 for Australia’s highest incidence cancers. The data is available by age, gender, socioeconomic status (SES) and location.

The findings confirm that outcomes for Australian cancer patients mirror those reported for other jurisdictions. The positive impact of national screening programs on detecting cancers at earlier stages is clearly evident.

Stage at diagnosis

  • Over 75% of newly diagnosed cases of breast and prostate cancers and melanomas were stage 1 or 2 cancers.
  • The proportion of colorectal cancers diagnosed at more advanced stages was higher in people aged <50 years than those 50 years and over.
  • In 42% of newly diagnosed cases of lung cancer the disease had already metastasized (stage 4). Only 18% were diagnosed at early stages 1 or 2.
  • For breast cancer, people aged 50 years and over had a higher proportion of stage 1 cancers at diagnosis than those aged less than 50 years.
  • For lung cancer, melanoma, and prostate cancer, persons aged less than 50 years had a higher proportion of stage 1 cancers than persons aged 50 years and over.
  • The proportion of cancers diagnosed at early stages was lower in remote and very remote areas compared to major cities and regional areas.

Survival by stage at diagnosis

  • For metastatic cancers (stage 4), survival was higher among people living in major cities compared to regional areas at 1 year (20% compared to 16%), 3 years (6% compared to 4%) and 5 years from diagnosis (3% compared to 2%).
  • For metastatic cancers (stage 4), survival was higher among people living in the highest SES areas (SES5, 25%) compared to lower SES areas (SES1-3, 17%) at 1 year from diagnosis.

Males had lower survival by stage at diagnosis than females

  • For stage 1 cancers at 5 years from diagnosis, survival for males was 62% compared to 75% for females.
  • For stage 3 cancers at 1 year (survival for males 54% compared to females 64%), 3 years (21% compared to 32%) and 5 years from diagnosis (14% compared to 22%).
  • For stage 4 cancers at 1 year from diagnosis (18% compared to 22%).

Survival progressively decreased with increasing time from diagnosis

Stage 1 cancers survival was 91% at 1 year, 76% at 3 years and 68% at 5 years.
Stage 2 cancers survival  was 70% at 1 year, 42% at 3 years and 32% at 5 years.
Stage 3 cancers survival was 58% at 1 year, 25% at 3 years and 17% at 5 years.
Stage 4 cancers survival was 19% at 1 year, 5% at 3 years and 3% at 5 years.

Source: https://ncci.canceraustralia.gov.au/outcomes;

Cancer Australia’s Stage, Treatment and Recurrence (STaR) projectCancer Australia, the state and territory population based cancer registries, the Australasian Association of Cancer Registries, and Australian Institute of Health and Welfare have collaborated to collect and combine data on incidence by stage at diagnosis data with mortality data from the National Death Index (NDI), for the top five high incidence cancers.

Who is subsidising who?

Over 30% of all prescriptions written in Australia for PBS-listed medicines are self-funded as they cost less than the relevant Co-payment (under co-pay).

General patients (co-pay $40.30, 1 Jan 2019) are the sole recipients of this cost-shifting as the dispensing fee ($7.29) is greater than the Concessional co-pay amount ($6.50).

So why do the 8% of Services to General patients account for over 30% of Government PBS/RPBS expenditure?

The bar graph shows Benefit paid and differences in proportion of Services attributable to patient categories with and without inclusion of under co-payment numbers. Another way to present this, and as reported in the annual PBS expenditure & prescription tables, is as Average Government Benefit paid per Service.

The line graph shows the evolution of Average Benefit per Service by patient category over time. This graph is based on figures downloaded from Medicare Australia website (includes both Section 85 & Section 100). The total services (approximately 205 K) match the prescription numbers presented in the PBS Expenditure and Prescription Report for the same periods, which exclude under-co payment prescriptions (since reporting started in 2012-13).

Why is the Average Benefit per Service currently over five times higher for the General patient category compared to Concessional?

 

Is there a difference in the demographics of General and Concessional patients such that higher cost (F1) medicines are prescribed more frequently to General category patients?

Benefit paid per service ($) by ATC CV CNS Anti-Infectives GI Respiratory Oncology-Immuno Other
2017/2018 $ 15.24 $ 28.65 $ 149.68 $ 31.64 $ 46.16 $ 966.48 $ 62.98

In recent years, the peak and drop for average benefit paid per service for the General ordinary category mimics that of usage of new Hepatitis C treatments. Is there disproportionate prescribing of these and other higher cost drugs, such as oncologics and immunomodulators, between categories? To test this, benefit figures were downloaded from the Medicare Australia website for Hepatitis C medicines and reveal a split between General and Concessional patient categories of 39% to 61% in $1.37 billion paid by Government during 2017-18. Total service numbers (72,266) show a similar ratio with General 37% and Concessional 63%. For the ImmunOncology medicine, Nivolumab (Opdivo, BMS) total services (40,560) were split General 32% to Concessional 65% in 2017-18.

Perhaps the cause is an artefact of the huge volume, relative to General patients, of low cost (F2) items prescribed, and paid by Government for Concessional patients? Rosuvastatin and Atorvastatin had the highest prescription volumes in 2017-18, of these 37% and 30%, respectively were under co-payment prescriptions.

As a proportion of over co-payment prescriptions in 2017-18, Safety Net (SN) accounted for by 1.3% of the General patient category and 18.2% Concessional. The addition of approximately $40 and $6.50, respectively per script to Government expenditure is unlikely to be responsible for the large difference in average benefit per service paid, given the General SN category is so small.

Suggestions on what may be driving this difference are welcome. Hope you can provide me with a D’oh moment!

 

Sources: Simpsons; Medicare Australia Statistics; PBS expenditure prescriptions report tables

Three slices of the PBS

Like it or not, product and service providers to the PBS are in competition. 

The winners and losers during development and roll-out of the 2015 PBS Access and Sustainability Package (PASP) left no doubt about that. 

A collaborative approach within the sector, to increase the size of the currently diminishing pie, will be more sustainable and deliver better health outcomes.

(1) Manufacturers

Manufacturer revenues accounted for approximately 70% of total Government Expenditure on the PBS/RPBS in 2017-18. In terms of costs, around 40% goes to innovator companies with single brand medicines in formulary F1; 20% to suppliers of multi-branded medicines in F2; and the remaining 10% for Combination products. The overall split, volume-wise, of prescriptions by formulary was F1, 11%; F2, 85% (half under co-payment and not represented in chart below); and Other, including combinations, 4%.

(2) Wholesalers

Wholesalers are remunerated via the regulated mark-up on ex-manufacturer price, currently 7.52%. This is agreed as part of the 5-year Community Pharmacy Agreement (CPA) negotiated with the Government of the day by the Pharmacy Guild. In 2017-18, payments to wholesalers represented only 4% of total Government expenditure on the schemes.

In their 2016 Submission to the King Review of pharmacy remuneration and regulation, the National Pharmaceutical Services Association (NPSA) re-iterated their view that the funding provided to wholesalers under the 6CPA is inadequate and unsustainable. This is even without including the impact of direct distribution to pharmacy model selected by some manufacturers.

The Community Service Obligation (almost $ 200 million per year) divided between full service wholesalers is not captured in the pie chart. However, a minimal increase on the previous CSO amount and a loss of  indexation during negotiation of the 6CPA, raises the question is anyone representing wholesalers at the table with Government?

(3) Pharmacy

The 2015 PASP/6CPA introduction of a flat, but CPI indexed Administration Handling and Infrastructure (AHI) fee successfully uncoupled community pharmacy remuneration from the price of medicines, and added to the growing range of professional services being remunerated. In addition to existing fees for dispensing, electronic prescriptions and incentives, such as to provide premium free medicines.

This approach has, and is, largely protecting pharmacy from the ongoing financial squeeze being experienced by manufacturers and wholesalers due to price disclosure, successive reforms and Department of Health activity.

Many of the professional service payments and pharmacy revenue, such as $600 m for the new Dose Administration Aid program and Minister Hunt announcements at APP2019, are not captured in the 26% of Government expenditure on the PBS/RPBS shown in the chart. For example, Clause 3.5 of the 6CPA notes: ‘The Commonwealth also estimates that community pharmacy will receive up to a further $4.8 billion from dispensing pharmaceutical items that are priced below the Maximum CoPayment.’ (1)

Bruce Annabel noted in a recent AJP article that ‘on average, pharmacies are receiving circa $30,000 pa services income’ with some generating over $200,000 pa.

With the PSA recently announced to join the 7CPA negotiations, and SHPA at APP2019 also wanting to play a role, as hospital pharmacists oversee more than 20% of annual PBS expenditure, there are going to be some unfamiliar faces at the table, very likely facing a new Minister of Health.

Sources: Department of Health Expenditure & Prescriptions Report; PharmaDispatch; Google Images

 

(1) The 6CPA bottom line of $18.9 billion to be paid to pharmacy over the life of the agreement, also excludes remuneration when community pharmacies dispense medicines under Section 100 special arrangements and the $372 m compounding fees which will be paid directly to chemotherapy compounders, who may not be approved suppliers.

The PBS pie

Using a pie analogy for Australia Federal Government spend on the Pharmaceutical and Repatriation Benefits Schemes in 2017-18, the impact of rebates is clear with approximately $1 in 5 paid out as benefits for medicines, ultimately returning to Treasury.

The total $2.36 billion in rebates was repaid wholly by innovator manufacturers on the basis of Deeds of Agreement with the Commonwealth. Such arrangements are necessary to enable PBS listing with a published price that globally protects the return of investment for capital risked while a new product is still in patent.

The size of the proportion is an indication of the negotiating power of the Department of Health on behalf of Australian taxpayers following a positive PBAC recommendation. It could also be considered a measure of why Australia is perceived as a ‘free rider’ in terms of investment in development. It may be a reflection of a distorted Health Technology Assessment process, with policies such as lowest cost comparator that often bear no semblance to clinical practice.

Graph Source: PBS expenditure prescription report, 2017-18. Image source.

PBS structural change

The long-term impact of 2007 PBS reforms are clearly evident when comparing patient categories by proportion of services and benefits. The introduction of separate formularies for single (F1) and multiple brand (F2) drugs plus successive price disclosure initiatives have fundamentally altered the structure of the Australian Pharmaceutical Benefits Scheme.

Despite ongoing increases in Concessional ordinary category (patient co-payment $6.50 in 2019) services (blue dashed line on graph), the cost paid/subsidised by Government (benefits, blue) has remained flat. Once a patient has paid $390 (60 items), the remaining services for the calendar year are provided free (Concessional Safety Net in green). Since the reforms the trend has been flat or downward, with recent significant drops most likely due to the $1 discount and other dampening policies on qualification.

While General category (co-payment $40.30 in 2019, shown in orange) services as a proportion of total has dropped away since the reforms, benefits paid by the Government for General patients (orange solid line) has almost doubled to account for 30% of total benefits in 2017-18. This is counter-intuitive as there are an increasing number of PBS items below co-payment and paid 100% out-of-pocket by patients. It is likely a consequence of the type and cost of medicines over the co-payment (refer earlier post PBS by therapeutic area showing increasing benefits paid for Oncology, Immunomodulators and Anti-infectives (Hep C) medicines).

The reducing service numbers in the General category (orange dashed) are influenced by the of loss of exclusivity on molecules with high prescription volumes and subsequent price disclosure. A similar dampening effect on the  Safety net qualification is also obvious over time.

Notes:

Patient category refers to a patient’s eligibility status at the time of supply of a PBS or RPBS pharmaceutical benefit. Concessions are available to Centrelink issued Pensioner Concession Card, Commonwealth Seniors Health Card and Health Care Card holders; as well as those with a Department of Veteran Affairs White, Gold, or Orange Card. General benefits apply if you do not have any of these cards. Further details are available on the PBS and DVA websites.

PBS Patient Category reports are available on the Government data.gov.au website as annual excel workbooks with month by month breakdown for the time period 1992 to 2016. Downloading directly from Medicare Statistics website is also possible for the period 1992 to current.

Image Source

PBS rewards innovation?

Last financial year (2017-18) the average benefit paid by the Australian Government per subsidised PBS/RPBS service was approximately AU $60 (1). However, separation of payments by therapeutic area (see graph) reveals a startling two-tier PBS with the majority of areas experiencing little growth in benefit paid since 1992-93, contrasted by those with consistent or dramatic spikes in growth.

How might this be interpreted?

 

One possibility is that, based on published list prices, the PBS/RPBS is rewarding innovation. The steady climb in benefit paid per service for Oncology-Immunological agents reflecting the ongoing innovation in this sector, with the consistent introduction of new, more cost-effective therapies.

The average benefit paid in 1992-93 for these drugs was $202.24 per service (in real 2017 dollars) representing approximately 3% of total cost to Government. In 2017-18, this had risen to $966.48 per service and Oncology-Immunology accounted for 32% of total Government cost of the PBS, but only 2% of total services. Specialists and patients in other therapeutic areas could question how representative this is.

In 2016-17, anti-infectives accounted for over 25% of all benefits paid by Government coinciding with peak uptake of newly listed Hepatitis C treatments. Although given the magnitude of rebates required in this area, revealed by a now-withdrawn poster, this information should only be considered guidance to trends.

The impact of price disclosure on benefit paid is clearly illustrated by the Cardiovascular (CV) therapeutic area. Over the 26-year period graphed, services as a proportion of all PBS activity have increased from 22% to 31%. Meanwhile, in 1992-93 the average benefit paid by Government per service was $35.39, peaking at $40.99 in 2003-04, and dropping to $15.24 in 2017-18 where CV medicines account for only 8% of benefits paid.

In terms of proportions, CNS medicines have remained relatively stable over the time period: 18% in 1992-93 to 22% in 2017-18 of services; 10% and 11%, respectively for benefits.

Gastro-Intestinal medicines have experienced a similar trend with an increase in proportion of services from 11% to 15%, while the benefit dropped overtime from 16% to 9%. This is most likely a reflection of the continuing prescribing of PPIs, and their loss of patent exclusivity.

The Respiratory therapeutic area has experienced a decrease in both the proportion of services (10% to 6%) and benefit (12% to 5%). A look at raw numbers of services shows a slight increase from 11.4 to 12 million over the period.

Methodology:

PBS/RPBS data downloaded from the Medicare Australia Statistics website for the period 1992 through 2018 was analysed by therapeutic area. The Anatomical Therapeutic Chemical (ATC) classification system, as recommended by the World Health Organisation (WHO) for drug utilisation monitoring and research (WHO 2019 Guidelines), was used. Benefit and service data was collated separately and then used to calculate the average Government benefit paid ($) per service by therapeutic area. Adjusting the benefit to adjust for inflation into real 2017 dollars did not alter the trends.

References: (1) http://www.pbs.gov.au/info/statistics/expenditure-prescriptions/expenditure-prescriptions-twelve-months-to-30-june-2018

Image source: http://cancer.nautil.us/article/196/why-cancer-drug-prices-keep-rising-in-the- us

The size of the PBS in 2030?

Extrapolating historical cost data, total Government expenditure on the PBS in 2030 is predicted to range from AU$11.5 to $14.5 billion.

The linear model based on actual PBS/RPBS expenditure after rebates in real terms (2017 $) from 1992-93 to 2017-18 (first graph) predicts a 2029-30 total cost of $14,400 m. This approach does incorporate the impact of Government policy changes and fiscal constraints, as well as the listing of new medicines and adjustment for inflation. Hence, the number is plausible assuming more of the same.

The 2% annual growth rate predicted by the trendline aligns with the recent IQVIA Institute Report on Use of Medicine (2019) that forecasts between 2019 and 2023, medicine spending growth rates in countries similar to Australia, are expected to be within a 1-4% range year on year. However, the Budget 2018-19 program expense for Pharmaceutical benefits, services and supply suggest that over $14 billion should be considered a stretch upside. This is due to the magnitude of projections for 2020-21 ($9,864 m) and 2021-22 ($9,787 m) once changes to supply chain arrangements take effect.

To better reflect the current environment, using only the past 10-years data, the 2030 PBS size lands at $9.7 billion (R2=0.1, low validity); while a trendline based on the last 15 years, predicts a total PBS of approximately $11.5 billion for 2030 (R2=0.7).

If benefit growth rates (%) are used rather than actual spend on the program as the basis for extrapolation, using linear extrapolation results in increasing negative rates (to -6.8%) and a PBS of AU$6.3 billion in 2029-30! As this is nonsensical, a moving average over 3 periods has been used to construct a trendline (second graph) and this provides an estimate of total PBS size of $11,865 m in 2030.

Using projected growth from 2014-15 to 2027-28 in real pharmaceutical spending per person from the most recent Intergenerational report (2015), plus the assumption that 80% of the growth is expected to come from non-demographic factors, the 2027-28 PBS total figure can be expected to be around $ 10.5 billion, supporting that the 1992-2018 linear extrapolation may be an overestimate.

Assuming the future holds the same in terms of policy & innovation, the total PBS will cost Government < $13 billion in 2030. How much of that amount goes to which parts of the sector will depend on how effectively stakeholders make their individual cases.

 

 

Sources: Expenditure data available from 1992 on the Medicare Australia Statistics website, picture & PBS logo

PBS Top 10

NPS MedicineWise has published* the Top 10 drugs for the financial year July 2017 to  June 2018 based on PBS and RPBS prescriptions. The data is based on date of supply of all prescriptions including those under the co-payment, sourced from the Department of Health in October 2018.

The top 10 medicines by cost to government are single originator brand products from the F1 formulary that, at the time of PBS listing, represented breakthrough innovations in their therapeutic area. Specifically, in the treatment of hepatitis C, macular degeneration and solid tumours. Overall, the ten represent 23% of government expenditure on the PBS and less than 1% of prescription volume.

Gilead’s Epclusa®, Harvoni® and Solvadi® occupy first, fourth and seventh place, respectively. These  account for almost 10% ($1.15 billion) of expenditure according to the data. However, rebates are not included and hence the real value that the Government place on this scientific breakthrough, on behalf of Australians, is unclear. The confidential nature of the effective price means it can be assumed that it is less than what governments and patients are prepared to pay elsewhere.

Anti-neoplastic and immunomodulating agents in the top 10 account for 7.3% of spend with AbbVie’s Humira®; Roche’s Herceptin® and the more recent checkpoint inhibitors, Opdivo® (BMS) and Keytruda® (MSD) represented. The top 10 for cost to government is rounded out by Bayer and Novartis with their macular degeneration treatments, Eylea® and Lucentis®, respectively. Xgeva® (Amgen) for bone mineral density is included at #8.

It must be assumed that risk share and/or special pricing arrangements are in place for all of these products. Thus the dollar values listed are likely to include rebate and cap re-payments that contribute to the PBS drug recoveries figure of $2.36 billion in the Department of Health Annual Report 2017-18.

Not surprisingly, the top 10 medicines by prescription volume all hail from the F2 formulary with multiple brands available. Cardiovascular treatments for cholesterol (rosuvastatin and atorvastatin) and blood pressure (perindopril) account for almost 14% of scripts dispensed. This is followed by protein pump inhibitors (esomeprazole and pantoprazole) for GORD and ulcers with 8% of volume. The  ‘overprescribed’ anti-infectives cefalexin, amoxicillin and amoxicillin with clavulanic acid, represented 7.5% of all prescriptions supplied in 2017-18. Metformin for Type 2 diabetes and escitalopram for depression and anxiety are # 8 and #10, respectively.

Source: Australian Prescriber* and Medicare Australia Statistics

The top 10 for defined daily dose (DDD) per thousand population per day is also included in the article and is suggested as a more useful measure of drug utilisation than prescription counts. It shows how many people in every thousand Australians are taking the standard dose of a drug every day. The list includes eight products for treatment of chronic cardiovascular disease (atorvastatin, rosuvastatin, perindopril, amlodipine, irbesartan, candesartan, telmisartan and  ramipril). Esomeprazole and metformin complete the ten.

*Australian Prescriber 2018;41:1943 Dec 2018 DOI: 10.18773/austprescr.2018.067

Image source

Health Data Snapshot, Part 3

A treasure trove of health data by location is available on-line. 

HealthStats NSW contains up-to-date information on over 200 health risk factors, diseases, locations, and specific populations within New South Wales. The flexibility and power of this resource is reflected in the snapshot of Hospitalisations by category of cause.  The blurry graph shows a time-series comparison by Primary Health Network and all NSW for ‘blood and autoimmune disorders’.  The data table is a click away below, the top down menu shows the other therapeutic categories that can be selected. Behind the drop-down are further choices based upon demographics, Aboriginal and Torres Strait Islander (ATSI) background and/or socioeconomic status.

The National Diabetes Supply Scheme provides an on-line mapping tool of registrants by various geographical locations. This can also be filtered further by diabetes type, age, gender, ATSI and socioeconomic status. The example shown is for the Local Government Area of Campbelltown in NSW where the estimated prevalence of diabetes is 7.3% of the population, which is classified as very high comparative to the national average rate of 5.3%.

 

In addition to these interactive, web-based applications, reports collating specific data are starting to be published. Most recently, the AIHW released Potentially preventable hospitalisations in Australia by small geographic areas.  This report provides information on 22 chronic, acute and vaccine-preventable conditions for which hospitalisation may have been avoided by the provision of timely and appropriate health care. Even though these reports usually contain data already available publicly, it is often cross-tabbed to provide different perspectives, and the quality of an AIHW analysis is assured.

 

As the MyHealthyCommunities website is to national and localised primary healthcare data, then the MyHospitals website is to hospitals data. The site, established by the AIHW in 2010, includes data for over 1,000 public and private hospitals nationally. Hospitals can be searched for by state or postcode. Individual hospital profiles, services, and performance indicators can be viewed and compared to other hospitals. Information can be presented as a data table, such as the Waiting Time in Emergency Departments by hospital type shown (above), or graphically such as the financial efficiency data (activity based funding) presented for The Alfred, Melbourne (below). The site includes an explanatory videos and notes to assist in interpretation of information provided.

Graphic source: http://fchd.org/Clinic/CountyHealthData.aspx