The aim of this study was to compare the cost/utility ratio of a management model of integrated, graded, intensive rehabilitation (GIR) versus usual care (UC) for patients with acquired Severe Brain Injury (SBI).

Materials and methods Study design This study was an organizational assessment of an integrated system of care (GIR) for patients with SBI and was designed as a prospective, randomised trial. An individual record was created for each patient with SBI; this record contained the patient's personal data, information prior to admission to intensive care or neurosurgery, clinical characteristics, the presence of devices, complications, time of hospitalisation in an Intensive Care Unit, and cognitive conditions. Once it had been determined that the study entry criteria were met, the patients were entered in the register and prospectively in parallel-group randomised to two different management strategies.

The concealment of allocation was guaranteed by the availability of beds in the ward. Patients and health care providers and assessors were blinded to the study designed to compare the strategies management. Data analysts were blinded to treatment.

Sample size The sample size requirements for this study are intended to provide adequate power for the analysis of the primary objective ( decrease event-related mortality). The simulation were run in statsoft six sigma software according to the following hypotheses: the mortality of severe brain injury used as a Null hypothesis was 58%. This percent value is expected to decrease at 40% in the study arm. The target type-I error(α) was 0.05. A sample size of 119 patients per group would allow a power of 80% to detect an individual risk score ratio of 0.70 with a type -I error of 0.05.

Based on the availability of beds in the ward, two groups were formed:

(i) the UC patients who, while meeting the criteria for access into the Graded Intensity of Rehabilitation Unit for patients with SBI (SBI-GIRU), did not enter the department because no beds were free and were, therefore, discharged from the Intensive Care Unit into the management models available, such as acute care, Intensive Brain Injury rehabilitation ( cod 75) brain injury rehabilitation (code 56) ( hospital admission, lasting on average 60 days, in which intensive rehabilitation care is delivered prior to transfer to another structure), long-term care, home.

(ii) The GIR patients who met the criteria for access into the SBI-GIRU and for whom a bed was available in the Unit, where an individualised programme of intensive rehabilitation care was structured.

The primary outcome for assessing the effectiveness of the two different forms of management was mortality; secondary outcomes included the improvement of quality of life evaluated using a classification system based on the Health Utilities Index Mark 2 (8) and the number of days of hospitalisation.

Patients Patients with acquired SBI referred to our unit between January 2007 and January 2010 were considered for this study. The diagnosis of acquired SBI was made on the basis of the clinical history, physical examination, clinical symptoms, Glasgow Coma Scale score <8 and neuro-radiological investigations.

After having been entered into the register, patients in the UC group began the monitoring phase, while the patients in the GIR group were submitted to clinical evaluation, prognostic stratification, and clinical and instrumental investigations, after which an individualised rehabilitation plan was made, characterized by structured multidisciplinary interventions that were periodically reviewed. The alternative management strategies are described in greater detail below.

Usual care In this case patients were managed according to the organization of the management model which took on the care of the patient. The organization of these models is characterized by one or two professional figures (physiatrists, neurologist), with hierarchical relationships, in spaces limited to a specific pathology; access is determined by clinical stability; the instruments of governance are guidelines and consensus and the rehabilitation programme is focused on functional and cognitive areas; the medical care process is governed by hierarchy. The technology in this model is limited to a specific specialty.

Graded intensive rehabilitation This is a result-oriented management model; it is multidisciplinary and has the purpose of determining the appropriateness of hospitalization and the patients' needs through the use of a ranking; the management is "disease-based", and the instruments of governance are the diagnostic-therapeutic rehabilitation pathways (DTRP), the quality system and product standards; the medical care process has separate strands with result-oriented autonomy; the technology present in this model is identified on the basis of health needs and support of vital signs.

Classifying patients In relation to their clinical conditions and complexity of care needs, three profiles of patients with SBI have been identified: the critically patient, the complex patient and the frail patient. The critically ill patient is a patient with abnormal vital parameters who requires constant monitoring and may need immediate therapeutic interventions. The complex patient is an individual with significant comorbidities who requires complicated, expensive and often invasive diagnostic investigations and/or numerous therapies sometimes conflicting with each other with a strong risk of interactions. The complexity of care increases in "frail patients". Frail patients have: (i) a marked susceptibility to develop acute illnesses that are manifested by atypical clinical pictures;(ii) rapid fluctuations of health, even in the same day, with a strong tendency to develop complications (failure cascade); (iii) a high risk of adverse events and iatrogenic events; (iv) slow and, almost always, only partial recovery; (v) continued requirements for medical interventions, need for ongoing assistance; and (vi) a high risk of mortality.

Levels of intensity of care On the basis of the above-described evaluation, the patient is assigned to the most appropriate of the following settings.

Level 1 (intensive): the patient is critical and unstable, dependent or potentially dependent on technological aids and at high risk of complications. Area of competence - red: monitored beds, single rooms, infusion pumps, telemetry, daily medical round, and intensive nursing activities 24 hours a day.

Level 2 (medium): the patient is clinically unstable and requires constant monitoring of vital parameters. Highly complex medical activities and intensive nursing care are prevalent. Area of competence - yellow: rooms with two beds, infusion pumps, telemetry, daily medical round, intensive, structured nursing activities, physiotherapy and speech therapy.

Level 3 (low): the patient is clinically stable and does not require constant monitoring of vital signs or the continuous presence of a doctor. Activities are prevalently physiotherapy and nursing care. Area of competence - green.

Level 4 (minimum): the patient is clinically stable and requires only social and nursing care for discharge. Area of competence - white.

The transition between different levels is determined by clinical criteria that define the area of competence required.

Staff The staff of the SBI-GIRU consists of a neurologist, an internist, a physiatrist, a cardiologist, a speech physician, twelve nurses with specific skills, a chief nurse, a nutritionist, two speech therapists, three physiotherapists, a psychologist and a social worker. The aim of the multidisciplinary approach is to facilitate recovery from disability caused by SBI.

Each member of the team has the duty to control, analyse and redesign the organization that supports the activities and processes set out in the individual rehabilitation plan. Three types of intervention - medical, nursing and physiotherapeutic - can be identified in the rehabilitation plan and are strongly integrated in order to achieve the rehabilitation objectives. The healthcare and rehabilitation processes are founded on guidelines and the criteria of evidence-based medicine.

The rehabilitation process The treatment plan is discussed in collaboration with the whole multidisciplinary team. An individual rehabilitation plan is drawn up which outlines: (i) the appropriateness of access; (ii) the type of access (SBI of traumatic, anoxic one, ischaemic or haemorrhagic nature); (iii) the objectives of the rehabilitation process (prognostic stratification, therapy optimization, prevention of secondary injury, management of devices, management of cognitive impairment, prevention of damage by immobility, recovery of motor function); (iv) use of the instruments to implement the plan (multi-disciplinary meeting, electroencephalogram, physiotherapy briefings, diagnostic radiology, consultancy, monitoring systems, infusion systems); (v) time-course of the plan; (vi) result indicators in the short to medium term; (vii) results indicators in the long-term; (viii) the method of discharge; and (ix) management counselling.

Follow-up A monitoring form was used for each patient to record the following information: personal data, outcome measures, employment issues, degree of socialisation, commitment of caregivers and the results of the Health Utilities Index Mark 2 questionnaire.

Outcomes Three types of outcomes were considered: management outcomes, functional results and a hard end-point (survival).

Management outcomes. The management outcomes considered were: access time, number of structures involved in the management of the patient, the healthcare path, the number of days spent in hospital, the type of management to which the patient was subsequently addressed.

Functional results. These included the number of devices at discharge, the Glasgow Outcome Scale (GOS) score, the Barthel Index score and any change in quality of life, assessed using the classification system based on the Health Utilities Index Mark 2 Survival. The patient's death related to the event index was considered as a hard event. In the survival analysis the two management models were considered as predictors of mortality. These predictors were adjusted for confounding variables (age, sex, aetiology, GOS score, Level of Cognitive Functioning score and indices of organ function). The association between independent predictors and outcome was assessed by comparing the follow-up data with a Cox proportional hazards regression model. The end-point of the study was event-related mortality. The distribution of the time function of the events was analysed with the Kaplan-Meier method

Cost analysis The cost/utility analysis was performed from the payer's prospective (the health authority). Hospital admissions for the implementation of the process and for health care - in particular, admissions to intensive care, acute care, rehabilitation and the social welfare sector - were considered. Management costs incurred at home were excluded. The prices used for calculations were those of the average daily cost of hospitalisation per type of management as listed in the Interregional Tariff Agreement [Tariffa Unica Convenzionale, TUC] 2009 for hospital healthcare services. The QALY was determined by calculating the years of life gained in management models adjusted for the quality of life. The difference in QALYs between the two models was obtained by considering the average increase of life derived from the survival curves (difference between the areas of the survival curves) adjusted for the quality of life.

Sensitivity analysis The sensitivity analysis was performed by calculating the cost/utility ratio using the extremes; in particular, changes in QALY and costs were evaluated by considering the extremes (confidence limits of the variations of the mean).

Statistical analysis To compare the clinical and functional parameters between the two groups we used the one-way analysis of variance for continuous variables and chi-square tests for categorical variables. For the analysis of prognostic indicators, the two management strategies were considered as independent predictors of outcome. The analysis was adjusted for confounding factors. event-related mortality was considered as an end-point of the study.

The association between independent predictors and outcome was assessed by comparing the follow-up data with a Cox proportional hazards regression model. The assumption of the proportionality of risk during the observation time, has been evaluated using the analysis of residues of Schoenfeld. Deaths from causes unrelated to the SBI were treated as censored observations. The distribution of the time function of the events was analysed using the method of Kaplan-Meier. The prognostic information is presented as relative risk with 95% confidence limits. Descriptive statistics are expressed as the mean value ± standard deviation. Probability values <0.05 are considered statistically significant.

Inclusion Criteria:

• Patients with acquired Severe brain injury
• patients with Glasgow Coma Scale score <8
• patients with Focal ischaemic lesion, cerebral contusion or hemorrhage on computed tomography or magnetic resonance imaging

Exclusion Criteria:

• patient with age< 18 years
• patients with previous degenerative nervous disease
• oncological patients with reduce life expectancy
• patients with impaired vital signs