Prepaid coordinated care for patients with diabetes: practices and patients bear the risks

General practices in Australia have no experience of, and no infrastructure for, managing prepaid funds for patient care. In the UK, the original fund-holding initiative was supported by a great deal of information and resources.3,4 In the United States, demonstration projects for patient-centred medical homes have clearly shown that additional resources should be paid, by capitation or other means, to enable practices to transform into centres for patient care coordination.17,18 This is necessary to give physicians time to coordinate care, and is akin to Medicare items for GP Management Plans and Team Care Arrangements (although these Medicare items do not assist with establishment and management costs). In New Zealand, prepaid funding has often been inadequate in terms of covering PHOs’ increased management costs. Supplementary funding has been required and, in some cases, PHOs have increased patient copayments to meet the increased costs.7 Although Australian health care services and delivery differ from those provided in these other settings, careful consideration must be given to the costs associated with managing prepaid funds and changing practice processes. The planned pilot projects need to determine the costs of implementing and operating the coordination of care for patients with diabetes.

The proposal appears to assume that the $950 paid to the registering practice is to provide all the health care services for the patient, not just for diabetes. If this is not the case, doctors will have to separate diabetes care from non-diabetes care. In general practice, where more than one problem is managed in a consultation,19 it is not realistic to expect a general practitioner to deal with the reasons for the encounter, partition the relative time and effort involved in each, bill Medicare for the non-diabetes portion, and assign the diabetes portion to the prepaid funds. Perhaps the prepaid funds are expected to pay only for a guideline-based set of specific services?14 This would imply that any extra services for diabetes, outside those specified in guidelines, could be billed to Medicare, possibly shifting costs back to Medicare and undermining the initiative. In both cases, justifying cost allocation across uncertain boundaries will add work and increase bureaucratic surveillance.

General practice’s raison d’être is whole-patient care. Diabetes may reach beyond the medical complications to, among other things, employment, relationships and psychiatric morbidity. GPs will find it difficult to determine where the health implications of diabetes end. For these reasons, prepaid funding should cover whole-patient care. This would make the accounting a simpler task, but would increase the likelihood of spending more than $950 per patient and create the risk of a reverse moral hazard, as practices may have to reduce services for needy patients if funding is inadequate.20 With the small amount of prepaid funding due to the relatively low number of patients registering in any one practice, this is a strong possibility.

As a basis for a fund to provide diabetes care, the payment of $950 is a blunt capitation measure. It is not sensitive to individual patient variation but relies on an average that might compensate a practice for servicing costs. The sum of registered patient payments will constitute a fund administered by the practice. A fund’s resilience depends on there being enough low usage patients to compensate for the fewer high usage patients. An indication of this risk has been provided by research on the probability of matching expenditure to fund size in the UK primary care setting. Using a sophisticated indexed capitation formula that is sensitive to local health indicators and therefore to likely expenditure, the probability of being 10% off target is nearly zero for a population of 500 000, but reducing the population to 10 000 increases the probability of the same outcome to 0.35.21,22

For a practice with a cohort of 300 registered diabetic patients and a resultant prepaid fund of $285 000, there may not be a sufficient reserve of low usage patients to compensate for the few patients who may require an unexpectedly high amount of care. The risk of being more than 10% off target is too high. A 10% overspend would bring a loss of $28 500, which could be enough to threaten the practice’s viability and patient services.

Practices will need to develop a process that enables patients to register voluntarily and may have to counter unrealistic patient expectations (related to costs) that have been engendered by the government’s promotion of the initiative.1 The latter is particularly important in view of the New Zealand experience, where increased patient copayments have been required in some PHOs.7 Building the register will take time, whether carried out opportunistically when patients attend or by using a recall system. It will also take time to build numbers of patients for alternative service delivery, such as group work. Practices will be faced with costs of registration, setting up alternative methods of service delivery and establishing systems for managing funds while waiting for funds to arrive.

A small decrease in patient satisfaction with UK fund-holding practices has been shown, which was partly related to the perception that doctors were concerned about keeping costs down.11 Levels of patient satisfaction with patient-centred medical homes in the US are not yet clear.18 Satisfaction was not studied in the Australian coordinated care trials.23 There is little information on patient satisfaction with New Zealand’s PHOs, although a national patient satisfaction survey has been developed for inpatient and outpatient services.24 Patient satisfaction is important to measure as it will be a patient’s own, and limited, health funding that his or her practice will spend.