Clinical outcomes associated with changes in a chronic disease treatment program in an Australian Aboriginal community

An epidemic of chronic disease has arisen among Indigenous people in remote areas of Australia, and renal failure is an important component.1-3 The underlying renal disease is marked by albuminuria, has multiple causes, progresses over time, and marks a greatly increased risk of death from non-renal as well as renal causes.4-7

In late 1995, we introduced a systematic chronic disease treatment program into an Aboriginal community with very high rates of renal failure and deaths from other causes.8 The first 3–4 years of operation of the program have been described previously.9-11 In 2000, responsibility for the program was assumed by the community’s Health Board, after a year-long handover. The Board had been constituted under the Coordinated Care Trials initiative and was funded for primary care services through the federal and territory governments. It retained the same protocols and treatments for chronic disease “activities”, and initially the same Aboriginal staff.

Here, we describe the subsequent evolution of the program and the clinical outcomes to mid-2003.

Treatment was offered to all adults with overt albuminuria (urine albumin–creatinine ratio [ACR] ≥ 34 g/mol), all with diabetes and microalbuminuria (ACR ≥ 3.4 g/mol), and all with blood pressure ≥ 140/90 mmHg. Treatment centred around use of the drug Coversyl (perindopril, a long-acting angiotensin-converting enzyme inhibitor; Servier Laboratories), blood pressure control, and attempts to control blood glucose and lipid levels where appropriate.9-11

Two hundred and sixty-four people were identified as eligible and enrolled in the program.10,11 Their mean age was 43.4 years; 53% were female, 44% had diabetes, 55% had hypertension, 64% had overt albuminuria, and 11% had serum creatinine levels above the laboratory’s reference range. In the first 3 years of the program, two-thirds of the participants reported that they were taking their medicines “some or most of the time”. In the treatment cohort, average blood pressure fell dramatically with treatment, and urine ACR, previously deteriorating, stabilised.5 Average serum creatinine level rose over the first 2 years, and then fell to below baseline values.

At an average of 3.4 years after enrolment, all-cause natural deaths in participants with overt albuminuria had fallen by 50%, and deaths from renal causes had fallen by 57%, when compared with historical controls matched for baseline disease severity.11 Whole-of-community trends in terminal events supported these estimates. Savings in dialysis avoided were estimated at $3.4 million, representing $12 879 per treated person over that period.12,13

There were problems generating lists of people due for rescreening and treatment, diagnoses, and clinical and laboratory results.14 Regular testing of the whole community decreased. Staff roles were reassigned, and the program coordinator left in late 2001 and was not replaced. The discrete chronic disease program was ultimately abandoned, and its activities were incorporated into mainstream care.15

In late 2001, a pharmacy review revealed that only 8% of people prescribed multiple medications were regularly picking up their weekly packs of medicine (Rollo Manning, consultant pharmacist to the Health Board, personal communication). Some concerted action followed, and, in mid to late 2002, medication orders to the supplier suggested that 35%–40% of people prescribed perindopril were taking it with some regularity.

The community’s Health Board went into receivership in September 2003. Responsibilities for health services were then assumed by the Northern Territory Department of Health and Community Services.

Participants in the treatment program were tracked through the community Health Board’s clinical information system and individual medical records, using their unique hospital record numbers. Clinical parameters, documented annually, were followed up to 31 December 2002.14,15 Deaths of program participants and in the community as a whole were documented through community and hospital records to 30 June 2003. All people who started long-term renal dialysis were documented from records of the dialysis unit in Darwin and, more recently, the satellite dialysis unit in the community itself.

A renal “death” was defined as the institution of long-term renal dialysis or death with end-stage renal disease without dialysis.6,7 Non-renal deaths comprised all other deaths from natural causes. Deaths from misadventure, intoxications and self-harm were excluded from the analyses, and participants were censored from the observations when these occurred.

A longitudinal analysis of the treatment cohort was undertaken. Use of contemporaneous controls was rejected, as it would have involved withholding best standard of care from some participants. Use of historical controls, as applied in our earlier analyses,10,11 was no longer possible, as the mean follow-up of the treatment cohort (with an extra 3 years) now vastly exceeded the follow-up of controls in the untreated state.

Average blood pressure levels and the proportions of participants with heavy albuminuria (ACR ≥ 100 g/mol) and raised serum creatinine levels (men, ≥ 120μmol/L; women, > 102 μmol/L) were calculated for annual intervals after enrolment. Because enrolment into the program was staggered, the numbers of people contributing to each observation fell as the interval since enrolment increased. Calculations were repeated for the 127 people who were followed up for 6 years or more after enrolment. Rates of terminal events were calculated as events/total person-years at risk. Cox regression analyses were used to calculate hazard ratios for terminal events in the remaining treatment cohort for the period 1 January 2002 to 30 June 2003, compared with terminal events in participants in the first 18 months after their enrolment in the systematic phase of the program. These hazard ratios were adjusted for sex, age and baseline urine ACR, which are strong predictors of terminal outcomes.6,7

Blood pressure and renal markers: Box 1A shows trends in blood pressure and renal function at annual intervals after enrolment in the participants remaining in the cohort. Average blood pressure began to rise 3 years after enrolment, followed by deteriorations in ACR and serum creatinine level 1–2 years later. Box 1B shows similar trends in the cohort of 127 people who were followed up for at least 6 years.

Renal and non-renal deaths: There were 49 renal and non-renal natural deaths in the treatment cohort over a total observation period of 1539 person-years: 20 people entered long-term dialysis, two died of terminal chronic renal failure without undergoing dialysis, and 27 died of non-renal natural causes.

Box 2 shows death rates over sequential intervals. For the first five intervals, renal death rates remained suppressed, at a rate about 60% lower than in matched controls in the pre-treatment era.7,8,10,11 However, rates increased in the last two intervals (2001–2002 and 2002–2003). Trends in non-renal natural deaths were not so clear, although the rate in the last interval was the highest in the observation period.

Box 3 shows that the risk of all-cause natural deaths (renal plus non-renal) in people remaining in the treatment cohort in 2002–2003 was 4.3 times the risk within the first 18 months of starting treatment. The risk of renal death was increased three-fold and the risk for non-renal death was increased 9.5-fold.

Box 4 shows the numbers of renal and non-renal deaths in the community as a whole, regardless of participation in the treatment program. It reflects the improvement during the systematic phases of the program, and the deterioration in the last two observation intervals. Almost 75% of the renal deaths after the handover of the program were among people in the treatment cohort. Trends in non-renal deaths were not definitive.

Eighty per cent of the people in the community who developed renal failure entered the dialysis program. Thus, the reduction in terminal events during the more active phase of the treatment program and the resurgence in the last few years had major impacts on the numbers of people in the end-stage renal disease treatment program and the costs of running it.

The events reported here, together with our experiences in a chronic disease outreach program in five other remote settings, suggest some ingredients that contribute to a successful and sustained program.16,17 There needs to be a commitment to ongoing chronic disease surveillance, treatment and education, as an essential and major element of primary care. Activities must be adequately and reliably funded and staffed over the long term. When health care funding and staff are spread thinly, activities perceived to be elective (eg, health promotion, mental health and chronic disease care) are the first to be pared in the face of further constraints or crises. An informed and supportive administration and a specifically trained Indigenous work force, which is allocated significant roles, respected and properly paid, are also essential.16-21 The clinical environment should be welcoming and respectful, regardless of the extent to which people can comply with lifestyle and treatment recommendations.

In addition, treatment decisions should be simplified and expedited,22,23 using a restricted number of drugs in each class, once-a-day dosing whenever possible, and easily transportable multidrug blister packs. Home visits and reminder systems based on daily events could improve medicine taking. However, there is serious need for innovation, such as a range of therapeutic dose “polypills”24 and transdermal drug delivery systems.

It is important to keep abreast of treatment advances. Better outcomes can now be expected from target blood pressures ≤ 125/75 mmHg in people with albuminuria, combination treatment with angiotensin-converting enzyme inhibitors and angiotensin-blockers,25,26 vigorous lipid-lowering therapy, better glucose control (often with added long-acting insulin), and more widespread use of aspirin.

Finally, a good clinical information system is needed to schedule visits, to follow clinical and laboratory results, to document process measures, outcomes and trends, to create community health profiles, to estimate cost effectiveness, and to assess the need for program modifications.13,21,27