The Program involved three teaching hospitals (Royal Brisbane and Women’s [900 beds], Princess Alexandra [640 beds] and Queen Elizabeth II [165 beds]) and five Divisions of General Practice (Brisbane North, Brisbane Inner South, Brisbane Southside Central, Bayside, and Logan). The consortium undertook a 2-year quality improvement program coordinated by a multidisciplinary steering group of leading clinicians and supported by district health service managers, divisional boards, and senior executives of Queensland Health. Five working groups, each with hospital and general practice representation, undertook the following program tasks: decision support, performance measurement and feedback, clinical information systems, patient self-management (assisted by patient representation), and hospital–general practice integration.

Consecutive patients admitted to any of the three hospitals between 1 October 2000 and 31 August 2002, and who met the case definitions (Box 1), entered the program. Patients registered between 1 October 2000 and 17 April 2001 comprised the routine care (or baseline) group; those registered on or after 18 April 2001 and whose care involved the quality improvement program were the intervention group. Quality of in-hospital care was assessed in all patients; quality of after-hospital care was assessed in a subset of patients who met eligibility criteria (Box 2) and provided written informed consent.

The targeted professionals were all consultant physicians, medical registrars and residents, nurses, clinical pharmacists and allied health professionals (n ≈ 450) working in cardiology, general medicine and emergency departments of participating hospitals; and all GPs (n ≈ 1080) who were members of the participating Divisions.

• Locally endorsed, evidence-based, clinical-practice guidelines disseminated to hospital clinicians and GPs as full-text guidelines, flow-chart pocket cards, laminated desktop guides, wall posters, and computer-based formats.

• Guideline-based reminders such as management checklists attached to hospital observation charts and chart stickers invoking consideration of specific treatments in eligible patients.

• Clinical pathways for chest pain assessment and management in the emergency department.

• Discharge summaries and practice evaluation forms issued to GPs, which acted to reinforce key management recommendations.

• Periodic educational presentations (eg, “grand rounds”, seminars, and workshops), hospital newsletter articles (eg, hospital drug and therapeutics bulletins, and divisional newsletters), case-based unit and departmental meetings.

• Academic detailing of hospital clinicians by trained clinical pharmacists, and of GPs, individually or in small groups (about 4–5), by trained detailers with no links to pharmaceutical companies.

• Feedback to target clinicians of quality indicators (see below) for in-hospital care every 6 months, and for after-hospital care at 8, 16 and 24 months after program commencement.

• Feedback of personalised quality indicators to individual in-hospital consultant units.

• Provision of patient self-management kits, comprising educational booklets (with treatment-specific risk-reduction estimates), patient diaries, individualised management plans, and self-explanatory medication lists.

• Pharmacist counselling of individuals about lifestyle changes, drug therapy, and risk-factor modification.

• After-discharge telephone follow-up by clinical pharmacists of high-risk patients (more than five classes of drugs, admission within previous 6 months, non-compliance, multiple comorbidities, age > 75 years, social isolation, risk of drug–drug interactions).

• Standardised discharge referral summaries with individualised treatment targets, advice on drug indications and monitoring, and follow-up arrangements.

• Copies of medication lists forwarded to patients’ community pharmacists at discharge, with telephone call from hospital pharmacist in selected cases.

• Telephone prompts to consenting patients to undertake follow-up visits to GPs.

Evidence-based, process-of-care indicators for in-hospital and after-hospital care were developed by consensus, as described elsewhere.2,3,9 In-hospital indicators were reported as the proportions of highly eligible patients (definite indication; no contraindication) who received specific care processes. Defining the indicator using eligibility criteria circumvented criticism from treating clinicians that indicators reported for their patients had not been adjusted for variation in casemix or risk. After-hospital care indicators were reported as overall rates, as detailed eligibility data could not be collected. In-hospital care data were abstracted retrospectively from patient records by trained nurses; after-hospital care data were obtained from GPs using standardised forms issued at 3, 6 and 12 months after discharge.

In-hospital care (Boxes 3, 4): Compared with the routine-care group, significant increases were seen in the number of intervention patients with ACS receiving the following: timely electrocardiography after emergency department presentation; angiotensin-converting enzyme (ACE) inhibitors and lipid-lowering agents at discharge; inpatient cardiac counselling; and referral to outpatient cardiac rehabilitation.

For patients with CHF, significant increases were seen in the numbers of intervention patients receiving the following: assessment for reversible, acute precipitants of cardiac decompensation; prophylaxis for deep venous thrombosis; requests for thyroid function tests (in those with atrial fibrillation); β-blocker prescription at discharge; and a recorded review of patient medications by clinical pharmacists. There was a trend towards increased use of imaging to assess left ventricular function.

After-hospital care: Relative to usual-care patients, more intervention patients with ACS were prescribed aspirin at 3 months (89% v 82%; P = 0.05), and, among those prescribed aspirin and β-blockers at discharge, more continued to receive these drugs at 3 months (92% v 84% [P = 0.03] and 85% v 76% [P = 0.05], respectively). For patients with CHF, more intervention than usual-care patients received β-blockers at 3 months (60% v 39%; P = 0.01) and at 6 months received ACE inhibitors (85% v 70%; P = 0.04); achieved ideal blood pressure levels (68% v 44%; P = 0.01); and were monitored for weight and salt/fluid intake (57% v 44%; P = 0.02).

Risk-adjusted mortality rates for patients with ACS trended downward from 5.0% to 3.8% at 30 days from 9.8% to 7.4% (P = 0.06) at 6 months and from 13.4% to 10.1% (P = 0.06) at 12 months. More notable reductions were seen for patients with CHF — 8.2% to 5.3% (P = 0.04), 22.8% to 15.2% (P < 0.001) and 32.8% to 22.4% (P = 0.005), respectively.

The median length of hospital stay for patients with ACS reduced by one day (from 7.0 to 6.0 days; P = 0.01), while that for patients with CHF remained unchanged (7.0 days). Same-cause readmission rates were not statistically different: 7.4% v 4.9% in ACS; 5.0% v 5.8% in CHF.

Our results may be subject to bias, as the study was unblinded, used historical controls, and achieved follow-up of after-hospital care in only 50% of registered patients. However, the mortality end-points at 6 and 12 months were ascertained for all patients and were risk-adjusted; in-hospital data were obtained for the entire cohort; and the total intervention period of less than 18 months minimised the effects of secular trends.

It is impossible to attribute specific process-of-care changes to specific QIIs within a multifaceted program. Systematic reviews reveal no singularly successful QII, and instead recommend deployment of multiple interventions.8

Processes of care directly controlled by individual clinicians (eg, drug prescribing) are readily amenable to behavioural change. Those more dependent on multilevel interactions within care systems (eg, emergency department triage or treatment responses, and timely access to complex modalities such as stress testing and cardiac rehabilitation) require interdisciplinary collaboration and/or work practice redesign, and hence change more slowly.

Our experience suggests the following pre-requisites: identification of problematic areas of care from practice audits; multidisciplinary quality improvement teams; multifaceted change strategies; ongoing performance feedback; and support of both managerial and clinical leaders. Various inhibiting factors will need to be overcome using locally designed solutions (Box 5).

The methods of BCC-CSSP have been extended into multihospital collaborations throughout Queensland targeting cardiac care, renal medicine and acute stroke care under the auspices of the Collaboratives for Healthcare Improvement instigated by Queensland Health (details available at <http://www.health.qld.gov.au/chi/default.asp>).

We are currently collecting data on long-term outcomes, readmissions and hospital costs in an effort to assess return on investment of our program,12 and such analyses will be the topics of future reports.