First steps towards price transparency: comparability of online out-of-pocket tools from Australian private health funds

Price transparency is one potential approach to improving health care affordability and consumer autonomy, based on the assumption that empowered consumers can shop for the best value health care. However, this is not straightforward and there is a lack of evidence on what effective ‘price transparency’ looks like. This is partly because such efforts are relatively immature, and because of multiple complex pricing elements present in the provision of care with multiple providers (e.g. doctors, hospitals, laboratories) and payers (e.g. government, private insurer, patient out-of-pocket (OOP)) involved.¹ This assumption may also be problematic if patients do not have the time or qualifications to judge health care quality, or are not in an area where there is competition between healthcare providers.

Despite the uncertainty around the impact of price transparency, both the US and Australian governments are exploring policies to facilitate it. As of 1 January 2019, the US Centers for Medicare & Medicaid Services (CMS) required all hospitals to release an online list of standard charges as a ‘first step’ towards price transparency.^2,3 The US President’s recent executive order on improving price and quality transparency, included a proposal for healthcare providers and insurers to provide information about expected out-of-pocket costs to patients.⁴

Interestingly, in the Australian context, the absence of a legislative requirement has not limited the development of such price transparency tools within the private health insurance (PHI) sector. Members of these PHI funds, when admitted to hospital, will often face OOP costs. Within the private health care sector, specialists and other private health providers (general practitioners, physiotherapists etc.) are free to set their prices. This results in these OOP costs being highly variable.⁵ The cost of PHI cover has also been increasing, so PHI funds are eager to demonstrate to their members the full price of hospital admissions (i.e. what members are actually being insured against), as well as informing their members what reasonable OOP costs look like. With limited fanfare, the three largest funds all released price transparency tools into the public domain over the last 24 months: HCF in March 2017 (https://www.hcf.com.au/preparing-for-hospital/cost-indicator-links, accessed 11 October 2019), Medibank in April 2018 (https://www.medibank.com.au/livebetter/what-is-the-cost-of-my-procedure/, accessed 11 October 2019) and Bupa in October 2018 (https://www.bupa.com.au/health-insurance/surgery-cost-calculator, accessed 11 October 2019). The fourth largest health plan, NIB, also introduced a similar tool that is password protected and only accessible to NIB policy holders.

This study aimed to compare and contrast the cost information these tools provide, and consider the implications this has for price transparency within the Australian healthcare system. Mehrotra et al.⁶ described four goals for price transparency: to do right by patients; to facilitate price shopping; to lift the veil; and to help providers ensure patients can afford care. We considered whether the online tools could achieve these goals for members within each respective fund and for any individual (i.e. system-wide transparency).

This is a well-timed review given the federal government’s recent announcement of a national strategy to target specialist OOP costs, following the formation of a Ministerial Advisory Committee on Out-of-Pocket Costs, chaired by Australia’s Chief Medical Officer, and the recommendation to build a website listing individual provider fees.⁷

The three funds each have links on their websites for various common in-patient procedures. The following pages on vaginal delivery childbirth provide an example: HCF (https://www.hcf.com.au/cost-calculator?pid=10, accessed 15 October 2019), Medibank (https://www.medibank.com.au/livebetter/cost-of-child-birth-vaginal-delivery/, accessed 15 October 2019) and Bupa (https://www.bupa.com.au/health-insurance/surgery-cost-calculator/vaginal-childbirth-cost, accessed 15 October 2019). For each procedure, the funds have used their claim data in different ways to show some expected cost information based on several listed assumptions.

We collected data for all the procedures listed on the websites of three funds, namely HCF (46 procedure pages; https://www.hcf.com.au/preparing-for-hospital/cost-indicator-links, last accessed 11 October 2019), Medibank (45 procedure pages; https://www.medibank.com.au/livebetter/what-is-the-cost-of-my-procedure/, last accessed 11 October 2019) and Bupa (16 procedure pages; https://www.bupa.com.au/health-insurance/surgery-cost-calculator, last accessed 11 October 2019), on 18 February 2019. Together, HCF, Medibank and Bupa represented 64% (10%, 27% and 27% respectively) of all Australian health insurance policies in 2017.⁸ We used the Web Scraper Chrome extension (https://www.webscraper.io/, accessed 11 March 2019) to save the procedure name, description, cost, OOP cost, excess, the maximum expected OOP cost and any Medicare Benefits Schedule (MBS) items listed on the page. Box 1 provides the definitions of these elements. After collecting this data using the Web Scraper extension, we used RStudio (http://www.rstudio.com/, accessed 11 October 2019) to explore and present the quantitative information from the pages.

Often the funds used slightly different procedure labels for the same service. For example, HCF claims for ‘Prostate Resection (TURP)’ included MBS Items 37203 and 37206, whereas Bupa and Medibank both stated that the common MBS item for ‘TURP (transurethral resection of the prostate)’ and ‘Prostate surgery’ was just 37203. We allocated these pages the same procedure category if the pages listed at least one of the same MBS item numbers. Bupa and Medibank did not list MBS items for both vaginal and Caesarean childbirth, but we still allocated these pages to the same procedure category.

Each fund uses their own claims data to calculate the various statistics shown on their online price tools. Funds may have different member demographics and charges from healthcare providers. However, each of the funds’ online price tools provided and used different information, so their estimates are not directly comparable. We could not assess whether the estimates were different across funds because of the provided data and statistics, or because there were true differences between the funds.

One issue was that the funds referenced different MBS items when describing the same procedure. Multiple MBS items may relate to a single procedure because these items describe various approaches, complications or additional procedures to the main procedure. The HCF pages stated the MBS items they used to select claims data. The Bupa and Medibank pages gave ‘common MBS items to ask your doctor about’, but it is not clear whether the funds based their claims selection only on these MBS items. On average, HCF listed 1.44 more MBS items for their procedures than Bupa, and 2.08 more than Medibank (based on the 14 procedures that all funds had pages for).

Table 1 lists the procedures on each of the funds’ pages. Procedures were grouped together if they had overlapping MBS items. There are many cases where one fund has grouped different procedures together compared with the other funds, which will affect the cost information. For example, HCF has different pages and costs shown for abdominal, vaginal or laparoscopic hysterectomy; Medibank has not separated different hysterectomy approaches and listed one price for hysterectomy. Bupa did not include hysterectomy on its site.

Table 1.  Procedures listed on the funds’ online price tools
Procedures were grouped together in the table if they had overlapping Medicare Benefit Schedule (MBS) items listed. The description of the procedure is shown as it appears on the fund’s website (HCF, https://www.hcf.com.au/preparing-for-hospital/cost-indicator-links; Bupa, https://www.bupa.com.au/health-insurance/surgery-cost-calculator; Medibank, https://www.medibank.com.au/livebetter/what-is-the-cost-of-my-procedure/, all accessed 18 February 2019). ACL, anterior cruciate ligament; CABG, coronary artery bypass grafting; TAVI, transcatheter aortic valve implantation

The funds also selected different claims data to use on these pages. Medibank states that it excludes admissions where there were no medical OOP expenses, whereas it is not clear whether HCF and Bupa also take this step. Specialists will not always charge OOP expense. For example, 32% of Medibank claims had no OOP costs in the grommet insertion example in Table 2. If the other funds’ price tools include these cases, then the statistics on OOP costs may be lower than what patients should actually expect in instances where an OOP cost occurs.

Table 2.  Information listed on the online price transparency tools for grommet insertion (Medicare Benefit Schedule (MBS) Item 41632)
Information listed on HCF (https://www.hcf.com.au/cost-calculator?pid=21), Bupa (https://www.bupa.com.au/health-insurance/surgery-cost-calculator/ear-grommet-surgery-cost) and Medibank (https://www.medibank.com.au/livebetter/cost-of-grommets/) websites, all accessed on 18 February 2019. DRGs, diagnosis-related groups

HCF and Medibank state the period their claims data is from (which overlap somewhat), but Bupa states this data is ‘updated periodically’. All funds use claims data solely from their private ‘participating’ or network hospitals.

The funds also use different statistics to show cost information. Bupa and HCF use a mean of claims data, whereas Medibank displays a ‘median view of cost information’.

Table 2 compares the information provided by the funds on the only procedure, grommet insertion, where all three funds referred to the exact same MBS item. The mean procedure cost Bupa reported was A$2502, whereas HCF reported a cost of A$3343 (34% higher than Bupa’s reported cost). Medibank’s reported median procedure cost was A$2330. The OOP cost from Bupa was A$213 (and A$250 excess), whereas HCF’s OOP cost was $734 (with A$0 excess) and Medibank’s was A$320 (with A$0–500 excess ‘depending on your policy’).

Medibank and Bupa also provide some additional cost information on their pages about the highest ‘expected’ OOP costs for the procedure. For grommet insertion, the Bupa page states, ‘You shouldn't expect to be out-of-pocket more than $590’, whereas Medibank states ‘Out-of-pockets for this procedure typically don’t exceed $920’. Across the 16 overlapping Bupa and Medibank procedures, the Medibank estimates were a mean (± s.d.) A$503 ± 457 higher than the Bupa estimates (if the excess was included, then Medibank estimates were on average A$449 ± 479 higher).

Fig. 1 shows the cost information from these funds’ pages for eight selected example procedures (of a possible 15 procedures overlapping between the three funds’ websites). The variation in the MBS items, the type of cost information (mean vs median) and the selected claims data means that these prices are not comparable. For example, we cannot compare Bupa’s reported total cost of A$67 692 for a cardiac defibrillator procedure with Medibank’s A$52 670. However, these figures (and all the comparisons across these pages) demonstrate the variety of ways insurers can choose to display price information, and provide an insight into how unclear this could be for users of these pages. This obfuscation has clear price transparency implications for members within funds, as well as transparency across the system.

Fig. 1.  Cost information for a selected set of procedures from Bupa’s (https://www.bupa.com.au/health-insurance/surgery-cost-calculator, accessed 18 February 2019), Medibank’s (https://www.medibank.com.au/livebetter/what-is-the-cost-of-my-procedure/, accessed 18 February 2019) and HCF’s (https://www.hcf.com.au/preparing-for-hospital/cost-indicator-links, accessed 18 February 2019) online price tools. The procedure description from each page is shown on the left. The Medicare Benefit Schedule (MBS) items from each page are listed on the right (italicised). The bars show the total procedure cost (Bupa and HCF base theirs on a mean of claims data, whereas Medibank shows a median cost) as the lightest (and longest) shade. The out-of-pocket (OOP) costs are printed on each bar (with excess included) and presented as the darkest shaded section on the left side of each bar. Costs in parentheses show the maximum expected OOP costs for Bupa (‘You shouldn't expect to be out-of-pocket more than…’) and Medibank (‘Out-of-pockets for this procedure typically don’t exceed…’) and are presented as the medium-shaded section of each bar.

Awareness of high and variable OOP expenses for private health care is increasing in Australia, with both the federal government⁹ and the Royal Australasian College of Surgeons¹⁰ recently proposing solutions. Price transparency is one possible approach to addressing this problem, and the three largest health funds in Australia have made inroads in providing this information to their members. These online tools, developed in a regulatory vacuum, provide a significant learning opportunity on the price transparency challenges within Australia’s healthcare system.

Mehrotra et al.⁶ proposed the ideal goals and information for price transparency. In Table 3, we review the utility of the information in the funds’ online tools against these goals. Price transparency should mean that patients know how much their care is going to cost (‘to do right by patients’) and they can be an informed consumer (‘facilitate price shopping’). The online tools may help reduce bill shock and raise awareness of possible OOP costs, and when a quoted OOP cost is excessively high.⁵ However, OOP costs do vary between regions and providers, and patients require information that is more specific to regions or providers. Providing this specific information will be a crucial step for the future development of tools that can facilitate price shopping. Bupa does have another price tool on average medical OOP costs for individual MBS items that are state specific (see https://www.bupa.com.au/health-insurance/Out-of-pocket-tool, accessed 11 March 2019). HCF and some other insurance funds have also collaborated with healthshare (https://www.healthshare.com.au/directory/find-a-health-professional/, accessed 11 March 2019) to provide OOP cost data for individual specialists, although a fund membership number is required to access the OOP estimates. The information on these pages may also misrepresent actual OOP and procedure costs. Including different types of admissions in the claims data may affect the cost estimates, such as excluding some MBS item numbers or pooling various procedure approaches together (e.g. the abdominal, laparoscopic and vaginal hysterectomy approaches). A successful, national website on OOP costs will have to clearly describe the type of procedures included in any cost estimates.

Table 3.  How the funds’ online price tools measure up against goals for price transparency (adapted from Mehrotra et al.⁶)
Price transparency tools are available from HCF (https://www.hcf.com.au/preparing-for-hospital/cost-indicator-links), Bupa (https://www.bupa.com.au/health-insurance/surgery-cost-calculator) and Medibank (https://www.medibank.com.au/livebetter/what-is-the-cost-of-my-procedure/) websites, all accessed on 18 February 2019. GP, general practitioner; OOP, out-of-pocket; MBS, Medicare Benefit Schedule; PHI, Private health insurance

The funds’ online tools also gave different information using this claims data. Perhaps the most interesting difference was the highest ‘expected’ OOP cost described by both Medibank and Bupa. The Medibank OOP estimate was usually much higher, which, in practice, may mean Medibank members are less likely to query high OOP costs compared with Bupa members. Of course, these numbers come from different data and assumptions, but, importantly, members may use and interpret them similarly in practice.

Price transparency, specifically public lists of specialists’ fees, may also have adverse effects on costs. It has been demonstrated elsewhere that some specialists may increase their fee if they find out they are below average, especially if the public perception is that higher fees relate to clinical quality (notwithstanding, health care cost and quality are not necessarily linked¹¹). It is not clear whether the funds’ online tools will have or are having this effect on specialist fees, particularly because the information provided is procedure based as opposed to provider based. Implementation of a national, specialist-specific website will have to consider these possible adverse effects. It may be a necessary, albeit complex, step to show quality and outcome measures along with individual specialist fees.

Something not considered in this review is whether these tools are actually used by fund members and the public. Previous US research has shown that few people offered price transparency tools actually use the information.¹² Implementation of a national website should consider when and how patients select a specialist, and when this price information is most useful in the Australian system.

The funds’ development of these online price tools is commendable because there was a gap for patient-directed information on OOP costs. Any critiques in this article are likely due to the inherent complexity of estimating and sharing health care price information. We do recognise that these pages are most likely a ‘consumer engagement’ tool. On their websites, funds encourage members to contact them before their procedure to discuss their policy and likely OOP costs, and these online pages may be an effective strategy to initiate these conversations. These tools may also be a ‘public relations’ exercise, as the funds show the value they offer to their members (by giving the full costs of procedures in the private sector that members are insured against).

Funds also contract and negotiate with private hospitals on rates for accommodation, theatre fees and other costs, and these negotiated rates can vary quite significantly. Therefore, prices can be commercially sensitive information between funds or in these negotiations with hospitals. Although funds may want to share information and achieve the price transparency goals within their membership (or at least prevent their members’ ‘bill shock’), there is a perverse incentive to fail at price transparency across the system in order to hide their position against their competitors. There is clearly still a gap for the development of national and specific information on OOP healthcare costs for all patients, without these commercial constraints, because we are far from achieving the goals for price transparency.

Kelsey Chalmers and Adam G. Elshaug have received consulting fees from Private Healthcare Australia for a project on low-value care data analytics. Adam G. Elshaug receives salary support as the HCF Research Foundation Professorial Fellow and is a Ministerial appointee to the Australian Medicare Benefits Schedule (MBS) Review Taskforce, receives consulting and sitting fees from Cancer Australia and as a Director of the Board of the NSW Bureau of Health Information. Shaun Larkin was Managing Director and Chief Executive Officer of HCF from 2010 to 2017.