Care pathways models and clinical outcomes in Disorders of consciousness

Abstract Objective Patients with Disorders of consciousness, are persons with extremely low functioning levels and represent a challenge for health care systems due to their high needs of facilitating environmental factors. Despite a common Italian health care pathway for these patients, no studies have analyzed information on how each region have implemented it in its welfare system correlating data with patients’ clinical outcomes. Materials and Methods A multicenter observational pilot study was realized. Clinicians collected data on the care pathways of patients with Disorder of consciousness by asking 90 patients’ caregivers to complete an ad hoc questionnaire through a structured phone interview. Questionnaire consisted of three sections: sociodemographic data, description of the care pathway done by the patient, and caregiver evaluation of health services and information received. Results Seventy‐three patients were analyzed. Length of hospital stay was different across the health care models and it was associated with improvement in clinical diagnosis. In long‐term care units, the diagnosis at admission and the number of caregivers available for each patient (median value = 3) showed an indirect relationship with worsening probability in clinical outcome. Caregivers reported that communication with professionals (42%) and the answer to the need of information were the most critical points in the acute phase, whereas presence of Non‐Governmental Organizations (25%) and availability of psychologists for caregivers (21%) were often missing during long‐term care. The 65% of caregivers reported they did not know the UN Convention on the Rights of Persons with Disabilities. Conclusion This study highlights relevant differences in analyzed models, despite a recommended national pathway of care. Future public health considerations and actions are needed to guarantee equity and standardization of the care process in all European countries.


| INTRODUCTION
Patients in vegetative state (VS) or in minimally conscious state (MCS), generally grouped in the term Disorders of Consciousness (DOC), are clinically classified as unconscious or low-responsive patients, respectively, and unable, or only partially able, to communicate their feelings and experiences (Bernat, 2006). In Italy, the Italian Ministry of Health technical report on patients with DOC (Stato Vegetativo e di Minima Coscienza -Epidemiologia, evidenze scientifiche e modelli assistenziali, http://www.salute.gov.it/imgs/C_17_pubblicazioni_1378_allegato. pdf, Accessed June 20, 2016.) reported rates of incidence and prevalence of 0.5-4/100.000 and 0.6-10/100.000, respectively, although these numbers seem to underestimate the real increasing incidence of VS and MCS in Italy as reported in the same document. Moreover, the incidence and prevalence of DOC is increasing for several epidemiological reasons, for example, aging of population (associated with the incidence of severe vascular brain injury) and the improvement in clinical management that determines an increase in the survival rate (Donis & Kraftner, 2011;Higashi et al., 1977;Lavrijsen, van   Materials and Methods: A multicenter observational pilot study was realized. Clinicians collected data on the care pathways of patients with Disorder of consciousness by asking 90 patients' caregivers to complete an ad hoc questionnaire through a structured phone interview. Questionnaire consisted of three sections: sociodemographic data, description of the care pathway done by the patient, and caregiver evaluation of health services and information received. Results: Seventy-three patients were analyzed. Length of hospital stay was different across the health care models and it was associated with improvement in clinical diagnosis. In long-term care units, the diagnosis at admission and the number of caregivers available for each patient (median value = 3) showed an indirect relationship with worsening probability in clinical outcome. Caregivers reported that communication with professionals (42%) and the answer to the need of information were the most critical points in the acute phase, whereas presence of Non-Governmental Organizations (25%) and availability of psychologists for caregivers (21%) were often missing during long-term care. The 65% of caregivers reported they did not know the UN Convention on the Rights of Persons with Disabilities.  . To the best of our knowledge, no studies have systematically analyzed this effect for patients in VS and MCS at the moment, and information on relationships between care pathways and clinical outcomes are still lacking.

Conclusion
This theme is also particularly important for ethical and legal issues related to the management of persons with DOC also considering the knowledge and role of caregivers during the care process of persons unable of self-determination.
Therefore, the aim of the present multicentric study was twofold: first, it aimed to analyze current NPCM-DOC implementation within different regional healthcare models focusing on the relationship between characteristics of care process with patients' clinical outcomes; second, to identify caregivers' opinions for each phase of the care process paying attention to critical points they noticed.

| MATERIALS AND METHODS
A multicenter observational study involving 10 intensive rehabilitation centers in different Italian Regions was conducted between November 2012 and November 2014. This study on care pathways for persons with disorders of consciousness was approved by coordinator Ethical Committee and was performed in accordance with the Declaration of Helsinki. Written informed consent was obtained from all caregivers' legal representative of all patients.

| Procedure
From October 2013 to May 2014, clinicians from each participating rehabilitation center contacted the main informal caregivers of all patients discharged with a diagnosis of DOC after traumatic or nontraumatic acute event (evaluated according to the Aspen criteria Giacino et al. (1977)  During phone calls, researchers asked the caregivers if he/she wanted to participate in the INCARICO project completing a structured phone interview in order to collect data on the care pathways of patients with DOC and collecting his/her opinions on the services provided to the patient he/she care for (main informal caregiver was defined highlighting the concept of assuming responsibility for the person who needs help as specified by Gould (2004)). If they accepted, researchers sent them the written informed consent and when caregivers sent it back to clinicians, professionals called again the caregivers and began the phone interview.
An ad hoc questionnaire (see Supporting Information) was developed for the INCARICO Project-phone interview. The questionnaire was developed considering data and results from national reports derived from previous researches (Giovannetti, Cerniauskaite, Leonardi, Sattin, & Covelli, 2015). It was developed taking into consideration NGOs frequently asked questions collected by caregivers on care pathways (Libro bianco sugli Stati Vegetativi e di Minima Coscienza. Il punto di vista delle Associazioni che rappresentano i familiari, http:// www.salute.gov.it/imgs/C_17_pubblicazioni_1377_allegato.pdf, Accessed April 25, 2017). If a caregiver did not respond to a question, or said "I don't know", interviewer proposed different issues related to services, professionals, and personal judgments for each clinical pathway phase. The "Not applicable" response was also possible when the service reported in the item was not provided for patients during the care process. If a patient died before the interview, caregiver could choose to complete the questionnaire or not, considering their emotional state. All interviewers asked the caregiver to complete the questionnaire after consulting all medical records available for clinical variables (e.g., hospitalization dates, last diagnosis, complete name of the institutions, etc.). For patients at home, the last diagnosis was collected considering the last medical records reported by multidisciplinary teams who evaluated clinical status of each patient every year.
Main variables derived from the questionnaire and analyzed in this study were: Length of Stay (LOS), Transitions along Phases (TP), Changes in Diagnosis (CD) and Mortality, Adherence to regional pathway of care (APC), number of caregivers available for patient assistance (nCG). LOS was calculated by summing the number of days spent by a patient in intensive care unit (ICU) and rehabilitation units in order to know how long a patient was hospitalized in health care service centers, noting also the first admission in ICU and RH. In this study, the variable "LOS in SA+RH" was obtained considering LOS in subacute and postacute units. TP consists of the number of admissions to health centers for each patient during his care pathway. For example, the TP of a patient who was admitted in three intensive care units, two semiintensive care units, three rehabilitation units, and one long-term care center during his/her care process was 9. CD represents the number of changes in patient's diagnosis registered along the care process.
In particular, this variable indicates the improving or the worsening in clinical diagnosis during hospitalization. In this study, comparison between the diagnosis at discharge after the first hospitalization in a rehabilitation center and the final one (at the moment of the interview) was used as an outcome measure. APC was calculated considering the percentage of patients who completed their cure and care process in the same region of residence. In detail, for each region, the number of patients who lived in that region and completed their care process in the same region was calculated and the number of patients who were hospitalized in that region but with their home residence in another region was also considered. nCG is the number of persons reported by caregivers who participated in caring process of one patient.

| Statistical analysis
Data derived from INCARICO questionnaire were analyzed in accordance with the following statistical methods: nominal variables are presented as number or percentage, continuous variables are presented as mean and standard deviation (SD) or median and interquartile range (IR) or minimum-maximum range. Non-normal distribution of the continuous variable was verified using the Kolmogorov-Smirnov test and analysis of skewness and kurtosis. Bootstrap method was used and confidence intervals were obtained with the "BCA" (bias corrected and accelerated) variations at 1,000 resamples for total median values reported in Table 1.
Series of univariate binary logistic regression analysis were performed to test relationships between each variable described in the previous section with outcome. Analysis details are reported in Supporting Information. All data were analyzed using SPSS 18.0 software (SPSS Inc., Chicago, IL, USA).

| RESULTS
Ninety patients were discharged from the participating centers in 2011. Of these, 24 (26.7%) died before the phone call interview as reported by their caregivers but seven of them decided to complete the INCARICO questionnaire in any case (Fig. S1).
At the time of death, the majority of them (no. 17, 73.9%) were hospitalized in nursing home, except one in hospital, whereas 6 (26.1%) were at home. In Table 1, the percentage of dead patients across regions are reported. Survivors were mainly males (no. 54, 60%), mean age was 50.5 (±18.8) years and mean time from acute event was 42.7 (±19.4) months. Regarding the mean age of patients who were hospitalized, the youngest (median age 26, min 24-max 66 years old) were from the Lazio region, whereas the oldest ones (median age 69.5, min 52-max 82 years old) were from the Sicily region.
Forty-six patients (54.8%) had MCS diagnosis at the time of the interview, whereas 29 (34.5%) were in VS and 9 (10.7%) were emerged from MCS remaining with a severe disability. For six patients, caregivers did not report diagnosis (all patients were dead before the interview). Changes in diagnosis were found mainly during hospitalization in the rehabilitation centers where 29 patients (39.7% of 73 patients analyzed) improved from VS to MCS, and 8 (10.9%) from MCS to Severe disability diagnosis, whereas 7 (9.5%) patients had a change in their diagnosis during longterm phase. Data collected on acute and subacute phases are reported in Table 1. Fifteen (20%) patients were re-admitted in rehabilitation units after their return at home or hospitalization in nursing homes.
Results from multinomial logistic regression analysis for acute and rehabilitation phase are reported in Table 2. No statistically significant p-values were found for the general model which included all independent variables both with forced and backward stepwise methods (first and second analysis steps). Interaction between LOS in RH and time from acute event or age showed low R 2 values but statistically significant. LOS was statistically significant both in the first (model A) and the second model (model B); so, the probability of finding changes in odds ratio in clinical diagnosis were directly proportionally to number of days spent in subacute/rehabilitation units by patients, so more days corresponded to greater probability of an improvement in clinical status (change from VS to MCS or from it to Severe Disability) than a worsening. No other variables than LOS seemed statistically significant in predicting the probability of a clinical status improvement than a worsening in our study. Table 3 shows the results of multinomial regression analysis for longterm care phase. Model with all independent variables included showed a R 2 greater than those after backward stepwise methods. Diagnosis of MCS or severe disability after rehabilitation phase significantly increase the probability of finding an improvement in patient outcome/ diagnosis in long-term care center (or at home) rather than a worsening (odd ratio increase of 10.636 and 9.391 in the first (model C) and the second model (model D), respectively). In the same way, the number of persons who cared for the patients represented a significant variable in predicting clinical improvement (than worsening) independently of the fact whether the patients were admitted in nursing home or at home. Seventy caregivers of 73 patients participated in the last part of the interview (3 caregivers of 7, whose patients were dead before interview did not complete the last part of the questionnaire). Their mean age was 54.7 years (±11.8), 40 (57.2%) were female and mean time dedicated to patient assistance was 13.5 hr/day (±8.1) at the moment of the interview. The median number of caregivers per patients was 3 (IR 1) and main caregiver was usually a patients' relative (wife 22.9%, mother 21.4%, sons 20%, husband 14.3%, father 12.9%, brother/sister 2.9%), cohabitant (1.4%), or persons paid for caring (4.2%). Results on caregivers' opinions are reported in Table 4.
Fourteen caregivers did not complete questionnaire related to the long-term care phase either because patients died during postacute phase or because they were still hospitalized in rehabilitation centers.
The section relative to long-term care phase was completed by 56 caregivers (43 patients were at home, whereas 13 in nursing homes).
Regarding caregivers' knowledge of the UN Convention on the Rights of Persons with Disabilities, 65.7% of caregivers reported that they did not know it, and 7.1% said they knew it only partially. The national T A B L E 1 Differences in hospital Length of stay (acute and postacute phase), number of interhospital patient transfers, and data on adherence to regional care pathway among Italian regions analyzed

Italian Regions
No. of   sample, seemed to be done by few patients (only 20%). The fact that changes in diagnosis were mainly found during the time spent in rehabilitation units is in line with previous literature (Estraneo et al., 2014).
This point can be explained with different points of view: in rehabilitation, the frequency of assessment is higher than in long-term care units and so the probability to find an improvement is quite high also June 20, 2016), discharge is possible when the individualized rehabilitation program is completed (no variation in conscious state was shown, absence of severe respiratory failure, etc.). However, the application of clinical guidelines is very heterogeneous because they are subordinated to regional welfare policies that are very different across regions (e.g., some regions limited the hospitalization time in rehabilitation unit) and this could affect the LOS in RH value as well as the number of beds available in chronic facilities in each region.
Results on "changes in the diagnosis", reported in  . Moreover, these results should be interpreted considering that most of the patients were at home during their long-term phase in our pilot study, and so the possibility for the main caregiver to have a support from other persons was fundamental both for physical health as well as for avoiding a too heavy emotional burden, a common risk for caregivers as reported in several articles (Corallo et al., 2015;Giovannetti, Leonardi, Pagani, Sattin, & Raggi, 2013;Leonardi, Giovannetti, Pagani, Raggi, & Sattin, 2012).
In the second area, one of the main problems reported by patients' relatives was the need to have information on patients clinical status (Leonardi et al., 2012) needed to analyze this issue. Moreover, in the present paper, we used a semistructured interview methods for population survey. As known, this approach implies that not all information can be checked and authors have to consider some answer received as true for definition. For example, the number of caregivers available to care patients were reported by main informal caregivers and we have no data to verify this information in our study. Finally, this study considers few data, from only 10 regions, in order to offer a new perspective and a starting point useful to all European countries to analyze the relationships between public health data, rehabilitation models, and patients' outcomes.
However, future studies are needed including a monitoring with standardized clinical scale scores and a larger sample than the one involved for this pilot research in order to compare different models.

| CONCLUSIONS
The present pilot study highlights that there are differences in health care pathways models, despite the common national pathway of care for patients in VS and MCS. The role of days spent in rehabilitation units and the number of caregivers caring patients seems to be important variables for the relationship between health care pathways and clinical outcome although future public health considerations are needed. Moreover, caregivers reported needs to improve services supporting them during all the care process.

ACKNOWLEDGMENTS
Two federations of NGOs, formed by caregivers of persons with DOC, two Regional Health Authorities and one Bioethics University Centre also participated in the study. The authors are grateful to all of them and to all caregivers who contributed to the study. Co-founded by the Health Programme of the European Union-PATHWAYS Project Grant Agreement no. 663474.

DECLARATION OF INTEREST
The authors declare that they have no conflict of interest and the study gained the approval from the Fondazione IRCCS Istituto C.
Besta Ethics committee and there are no constraints on publishing.