Diabetes model improves outcomes but is not cost-effective
An evaluation of the three-year, $33 million Diabetes Care Project (DCP) showed that while it did see an improvement in blood glucose levels and other clinical indicators in patients in one arm of the study, the funding model used was not cost-effective and was unlikely to be able to be scaled up nationally.
The DCP began in July 2011 with funding from the federal and Victorian governments and was intended to measure a number of innovations, including the use of an electronic shared care planning tool, flexible funding models including capitation-style payments to GPs based on the idea of the medical or healthcare 'home', as well as the use of care facilitators working to coordinate patient care with a number of general and allied health practices.
The DCP had three arms: a control group that received business as usual care; Group 1, which tested improvements through the use of the cdmNet shared care planning tool from Precedence Health Care as well as continuous quality improvement processes; and Group 2, which tested those components along with flexible funding based on risk stratification, payments for quality and funding for care facilitation.
It was a big trial, involving 184 general practices in three states and 7781 people with diabetes, and its primary clinical endpoint was the difference in the change in HbA1c levels between treatment groups at the end of the project.
Secondary outcomes included changes in other biochemical and clinical metrics, incidence of diabetes-related complications, health-related quality of life, clinical depression, success of tailored care and and economic sustainability.
The evaluation of the trial found that for Group 2, there was a statistically significant improvement in HbA1c levels compared to the control group and clinically modest secondary outcomes, but that came at a higher cost. It cost $203 more per person per year than the control group, with increased costs in payments to GPs and allied health practitioners offset by reductions in potentially preventable hospitalisations.
However, the evaluation found that while the cost difference was not statistically significant, “it is unlikely that the particular funding model implemented in the DCP would be cost-effective if rolled out more broadly”.
Participants in Group 1 did not experience a significant improvement in HbA1c levels or other clinical metrics, aside from a small improvement in renal function. (Previous research had shown otherwise, with a smaller study involving Precedence Health Care showing significant improvements in clinical outcomes for patients on a general practice management plan created using cdmNet.)
“The DCP demonstrated that improved information technology and continuous quality improvement processes were not, on their own, sufficient to improve health outcomes,” the evaluation found. “However, combining these changes with a new funding model did make a significant difference.”
cdmNET had some additional functionality included for this project that could be harnessed in future, the report says, including patient registration, automated risk scoring, care planning and clinical protocols, provider bookings and eReferrals, care tracking, a common patient record, a patient portal, performance management and analytics, and billing management.
The study found that most of the information recording for the DCP was done by practice nurses and practice managers rather than the GP, and use of the patient portal was not high. Only seven per cent of patients in Group 1 and 18 per cent of patients in Group 2 accessed the portal.
However, the report found that cdmNET was an enabler of greater collaboration across care teams and was used much more frequently by Group 2 versus Group 1. This may have been due to the use of the tool as part of the flexible funding model. In Group 2, GP and allied health activity was tracked at the per-patient level so funding could be allocated to the provider.
“In Group 2, GPs used the tool twice as often, practice nurses used it three times as often, and [alled health] used it six times as often as their respective counterparts in Group 1,” the report found. “A large proportion of these encounters with the tool involved entering information that was shared across a patient’s multidisciplinary team.”
In addition to shared care planning using cdmNET, the DCP also investigated different funding models, including quality-based payments to GPs and the ability of GPs to allocate funding to an allied health practitioner such as a podiatrist.
The DCP actually grew out of a much bigger plan first announced by Labor's health minister Nicola Roxon in March 2010, who put aside funding of almost $450 million over four years to investigate the flexible delivery of services to people with diabetes by their general practices.
Both the AMA and RACGP kicked up a bit of a stink about this idea, including the concept of pay-for-performance targets for general practice. Ms Roxon subsequently reduced the scope of the original plan and decided to fund the pilot of the DCP instead.
In this trial, both GPs and allied health practitioners received funding on a tiered model according to risk. General practices received payments on a quarterly basis of between $130 and $350 per person per year.
These payments were paid on a population basis and were not tied to activity. While GPs did not claim MBS items for creating care plans and TCAs, they were able to continue to claim for standard consultations and other items.
Allied health funding remained tied to activity but the amount was tiered based and there was a broader range of consultations that could be funded, including short consults to see a podiatrist to get toenails cut and payments for phone consults.
There was also funding available to GPs for quality improvement based on improved clinical outcomes, improved clinical processes and patient experience.
Funding was also made available for care facilitators, who were responsible for the holistic care of participants and worked with a number of different general and allied health practices and on average over 280 patients each.
Their role involved reviewing patient data, booking case calls, scheduling home medicines reviews (HMR) or mental health reviews, or finding alternative allied health professionals for participants in the event of availability issues.
Care facilitators were also responsible for supporting and educating practices and participants to use cdmNet for patient registration, risk stratification and care planning.
Cost of care
As expected, there was a small percentage of patients who were responsible for most of the costs in the project. The evaluation found that the most costly five per cent of participants accounted for 62 per cent of potentially preventable hospital costs, 47 per cent of other hospital costs and 13 per cent of PBS and National Diabetes Services Scheme (NDSS) costs.
“The most costly five percent of participants had an average total cost of $48,623 per person per annum, compared to $16,560 for next 15 percent of the DCP cohort, and $4,670 for the remaining 80 percent of the DCP cohort,” the report says.
Costs for both Group 1 and Group 2 participants were higher than the control group – $718 and $203 per person per annum higher respectively – although this was not statistically significant. However, costs components such as GP and allied health rose, as did PBS costs and there was the new cost of the care facilitator.
Even when reduced hospital admissions and other acute care savings were taken into account, the evaluation found that the new funding arrangement was still not cost-effective.
“Measuring the cost-effectiveness of the Group 2 model of care is challenging because the main benefits of improved blood glucose control would be expected downstream in the form of reduced complications (i.e. outside of the 18-month trial period).
“These benefits were extrapolated (based on HbA1c and other clinical changes), along with costs, for an extended period … [but] overall, there is no evidence to suggest that the Group 2 model of care would be cost-effective if adopted for longer, with large uncertainties regarding both the net cost and benefits of the intervention.
“Were a scheme similar to the DCP to be rolled out more broadly, the funding model would need to be recalibrated to produce a greater likelihood of cost-effectiveness.”
The evaluation committee made three recommendations arising from the DCP:
- Change the current chronic disease care funding model to incorporate flexible funding for registration with a health care home, payment for quality and funding for care facilitation, targeting resources where they can realise the greatest benefit
- Continue to develop both eHealth and continuous quality improvement processes
- Better integrate primary and secondary care and reduce avoidable hospital costs. “It is possible to predict who is at risk of potentially preventable hospitalisations using data from the DCP.”
In response to the release of the evaluation, Health Minister Sussan Ley has decided to extend the consultation period for the promised National Diabetes Strategy until the end of the month to allow the public to review the report, along with the Australian Institute of Health and Welfare's (AIHW) Incidence of type 1 diabetes in Australia 2000-2013 report.
Posted in Australian eHealth