FACTORS AFFECTING LENGTH OF HOSPITAL STAY FOR PEOPLE WITH SPINAL CORD INJURIES AT KANOMBE MILITARY HOSPITAL, RWANDA

In patients with spinal cord injuries increased length ofhospital stay is often as a result of secondary complications such as pressuresores, urinary tract infection and respiratory infection. An increased lengthof hospital stay was observed at Kanombe Military Hospital in Rwanda.The aim of this study was to determine specific factors affecting length ofhospital stay for individuals with spinal cord injuries at Kanombe MilitaryHospital in Rwanda. The records of 124 individuals with spinal cordinjuries who were discharged from the hospital between 1st January1996and 31st December 2007 were reviewed to collect data. Information collected and captured on a data gathering sheetincluded demographic data, information relating to the injury, occurrence of medical complications and length ofhospital stay. Linear regression analysis was computed in SPSS to determine factors affecting the length of stay.The necessary ethical considerations were adhered to during the implementation of the study. Current employmentstatus and the occurrence of pressure sores were significantly associated with the length of hospital stay (p=0.021 andp=0.000 respectively). A strong relationship was noted between pressure sores and length of stay (R= 0.703). There is aneed for all members of the rehabilitation team to devise and implement effective measures to prevent the developmentof pressure sores, in patients with spinal cord injuries in the study setting.

medical complications throughout life.These may interfere with health and wellbeing, social activity, produtive employment and quality of life (Middleton et al 2004;Ackery et al 2004).Patients are often readmitted to hospital due to complications with the genitourinary, gastrointestinal, skin, musculoskeletal, neurological, respiratory and cardiovascular systems (Middleton et al 2004;Savic et al 2000).The occurrence of pressure sores is among the most common longterm secondary medical complication in persons with SCI (Garber and Rintala 2003).
The majority of patients with SCI receive hospital rehabilitation to address the impact of the injury on the per son (Joseph 2011;van Langeveld et al 2011, Henn et al 2012).The rehabilita tion of patients with SCI aims to enable the fullest range of activities and active participation in all aspects of human life (Chappell and Wirz 2003).The provi BACkgROuND Spinal Cord Injury (SCI) is a devastat ing condition that requires intensive and specialized clinical rehabilitation.It occurs most frequently to young, healthy individuals around the world (Post et al 2005;Ragnarrsson et al 2005).A SCI leaves individuals with a wide range of impairments, activity, limitations and participation restric tions (Kirchberger et al 2010).Common impairments experienced by indivi duals with a spinal cord injury include changes in emotional, sensory, defe cation, urinary, motor and sexual func tions.Individuals who have suffered a spinal cord injury could also have activity limitations and participa tion restrictions relating to self care, mobi lity, interpersonal relationships and community, social and civic life (Scheuringer et al 2010).
The devastating changes caused by SCI also predispose the individuals to the development of various secondary sion of services including rehabilitation is however challenging in developing countries (Kautzky and Tollman 2008) especially in the transitional period after a war where rehabilitation centers do not have adequate staff, facilities or equip ment to assist individuals who have spinal cord injuries (Adler et al 2004).
Length of stay following spinal cord injury varies.For instance the mean LOS for spinal cord injury ranges from, 5661 days in Australia (New et al 2002), and 91143 days in Italy (Celani et al 2001) 106239 in Israel (Ronen et al 2004), and in the Western Cape South Africa a mean of 73 days for patients admitted to a specialized inpatient faci lity in the Western Cape (Joseph 2011) and a mean of 95 days for those admit ted to a Private Rehabilitation Hospital (Henn et al 2012).There is limited published literature relating to the LOS in other African countries.Factors reported to be associated with LOS in hospital for patients after SCI are per sonal factors and hospital factors.The personal factors include age, severity of injury and the development of secon dary medical complications, notably urinary tract infections and pressure sores (Herm, et al 2000) whereas hos pital factors include being treated in a public or private hospital, the availabi lity of beds and the staffing levels (Simpson et al 2005).
While working as a physiotherapist in the physiotherapy department at Kanombe Military Hospital (KMH) between 2002 and 2006, the researcher noted that the majority of SCI are hos pitalized for many years.The motivation for the study was therefore to determine length of stay and factors affecting the length of hospital stay among spinal cord injured patients at this hospital as the patients are admitted for a much longer period than what is reported in the lite rature.This extended length of hospital stay at KMH results in increased medi cal costs (Wong et al 2011).

Research Setting
The study was conducted at the Kanombe Military Hospital in Rwanda.The spinal rehabilitation unit is located within this hospital.The hospital has a bed capa city of four hundred inpatients.It is an acute hospital with departments such as medicine, surgery and gyneco logy.Patients admitted to the spinal unit are managed by a multidisciplinary team comprising of physiotherapists, nurses, psychotherapists and doctors.The team comprised of two general practitioners, one physiotherapist, 8 nurses and two psychotherapists.

Study design and sample
This study utilized a quantitative retro spective study design.
All patient records of individuals who sustained a spinal cord injury, and who were dis was obtained.The researcher trained one physiotherapist who was not asso ciated with the health setting where the study took place as the research assist ant.Training of the researchassistant helped him understand the topic as well as the aims and ethical considerations of the study.Arrangements were made with the staff at the hospital for accessing the folders.After receiving all neces sary written documents authorizing the researcher and the research assistant to carry out the study in the Military Hospital, and adhering to all the neces sary ethical considerations, appoint ments were made with the archives officers for data extraction from patients' medical folders.

Data Analysis
Data were numerically coded and cap tured using the Statistical Package for Social Sciences version 15.0 (SPSS) software programme for Windows.Descriptive statistics of the data namely frequencies expressed as percentages were used to present information regard ing demographic characteristics of the data.Means as a measure of central tendencies were used to calculate the average age of the sample, and the length of stay (LOS).The chisquare test was used to identify whether there were any associations between demographic data and, occurrence of secondary medi cal complications with the length of stay.Alpha was set at 0.05.The linear regres sion analysis was used to determine fac tors affecting the length of hospital stay.

Ethical considerations
Ethical clearance was obtained from the relevant committees at the University of the Western Cape, South Africa.Permission was obtained from the Minister of Health in Rwanda as well as from the director of medical services in Rwandan Defense Force (RDF) in the Ministry of Defense.Further permis sion was obtained from the director of Kanombe Military Hospital where the study was conducted.To ensure ano nymity and confidentiality of the data, no names were documented during the capturing and analysis of the data.charged from Kanombe Spinal Cord Rehabilitation Unit (KSCRU) from 1 st January 1996 to 31 st December 2007 were perused and relevant information was extracted.The period 19962007 was chosen by the researcher as hospi tal records were not always kept up to date as a result of the war and genocide which occurred during 1994.

Instrumentation
A selfdeveloped datagathering sheet was used to collect the data.The researcher developed this instrument based on a previous study (Middleton et al., 2004).The final datagathering sheet was divided into sections A and B. Section A captured sociodemographic data such as age, gender, education level, employment status and the causes of SCI.The level of the lesion, which was determined clinically, was docu mented as either paraplegia or tetra plegia.Section B identified LOS and occurrence of secondary complications such as pressure sores, urinary tract infections, and respiratory infections which could have occurred at any stage during the patients admission to hospi tal.To achieve face and content validity, the data gathering sheet was sent to experts in the field of rehabilitation of SCI patients.These experts included a physiotherapist working at a high inten sity rehabilitation centre and a senior lecturer, both who had experience in the rehabilitation of patients with SCI.These experts were satisfied that the instrument would capture the data needed to address the aim of the study.To ensure the reliability of the instru ment, the researcher and research assis tant compared data collected from the same five patient records prior to the commencement of the study.Due to the small sample size no statistical tests were used to determine the level of reli ability.(Professor Madsen, Department of Statistics, University of the Western Cape, personal communication, 30 October, 2008).The consistency of the documented data was very high indicat ing a high level of interrater reliability.

Procedure of data collection
Prior to data collection, permission to conduct the study and ethical clearance

Demographic status of the participants
A total of 142 folders of patients with SCI admitted to Kanombe Spinal Cord Rehabilitation Unit (KSCRU) from 1 st January 1996 to 31 st December 2007 were accessed.The folders of 18 patients were excluded from the study due to missing data.The final sample there fore consisted of 124 participants.The information relating to the demographic status (gender, age, marital status, cur rent employment status and level of education attained) of the sample of 124 participants with SCI is presented in table 1.
The mean age of the sample was 32.2 years with a standard deviation of 6.31.The ages ranged from 19 to 58 years.The majority of the sample was young adult males (97.6%) and was under the age of 42 years.Pertaining to marital status, more than seventy percent were single compared to 27.4% that were married.

Factors relating to the SCI
The majority of the sample (79.8%) was paraplegics.The causes of the spinal cord injuries among the sample were many and included gunshot wounds (64.5%) as the major cause, followed by road traffic accidents (30.4%) falls from a height (4.1%) and other (1%).

Occurrences of secondary complications:
The occurrence of secondary compli cations among the sample included pressure sores, respiratory infections, urinary tract infections and pain.These secon dary medical complications are presented in table 2.
The vast majority (80.6%) of the sample developed pressure sores while 37.9% had urinary tract infections.Within the sample of 124 patients who were discharged from the hospital in the period 19962007, the majority (75.8%) had more than one secondary medical complication pressure sores were pre valent in conjunction with other compli cations as is presented in table 3.

length of hospital stay (lOS)
The length of stay (LOS) in the hospi tal was calculated from the first day of

Factors affecting lOS
The associations between demographic variables and secondary medical com plications with LOS were computed using cross tabulation and chisquare tests.A significant association was found between current employment sta tus and LOS (Pvalue =0.021) as well as between pressure sores and LOS (Pvalue=0.000).The linear regression analysis revealed a high positive corre lation between pressure sores and length of stay (R= 0.703).Furthermore, the analysis also revealed a moderate posi tive correlation between employment status and length of stay (R = 0.575).

DISCuSSION
The results of the current study show a high positive correlation between occur rence of pressure sores and length of stay (R= 0.703).The occurrence of pres sure sores in the participants could have been as a result of the limited number of medical personnel available to provide appropriate preventative care.Rwanda lost up to 6080% of its health profes sionals during the war either as a result of death or fleeing the country (Adler et al 2004).This workforce was still not recovered at the time the study was conducted and therefore limited staff was available to care for patients.The management of pressure sores includes preventative measures, medical interven tions and the use of correct equipment (Staas and Cioushi 1991).Preventative measures include correct position, pres sure relief, the correct use of equipment and observing the skin for signs of increased pressure.The high occurrence of pressure sores in participants could have been as a result of a lack of imple mentation of preventative measures or the patients not complying with educa tion given about pressure care.The development of pressure sores at any stage in the rehabilitation process affects the rate of progress during rehabilita tion, thereby prolonging their length of stay (Ash 2002).As was the case in the present study, Regan et al (2006) stated that pressure sores are a lifelong threatening secondary medical compli cation of a SCI which have the potential to interfere with physical wellbeing, and impact on the overall quality of life.
Although pressure sores are prevent able, these authors argue that their pres ence disrupt rehabilitation and prevent individuals with SCI from working and interfere with community integration.Spinal cord injury patients who develop a pressure sore at any stage between injury and discharge are in hospital for an average of 177 days which is 55days longer (95% CI 3378.days) (Ash 2002) The results of the current study also revealed a significant association between employment and LOS (Pvalue = 0.021) whereas other demographic data such as age, gender, level of edu cation and marital status revealed no significant association with length of hospital stay.The majority of the parti cipants were documented as being retired at the time of data collection.These participants were those who had been previously employed by the mili tary and had received severance pack ages.They included young individuals who had not been reemployed follow ing their injury.Despite Rwanda being a signatory of United Nations convention on the rights of persons with disabilities, and with good policies in place to pro mote equal participation of people with disabilities (UN, 2008), persons with disabilities still find it difficult to access jobs.The return to work rate post spi nal cord injury has been reported to be as low as 33% (van Velzen et al 2009).Factors that affect the number of people taking up employment following spinal cord injury include access to places of work, and employers' attitudes towards people with disabilities (Yasuda et al, 2002) as well as age at onset of injury, level of education and motor function (Hess et al 1999).In the current study the LOS could have been decreased if the participants could find employment and be discharged to a different setting as is discussed below.The army could still have been taking responsibility for them as they were injured in the war.
In the present study, LOS (mean = 6.56 years) was much longer when compared to what is reported in the literature.A study conducted in the Netherlands reported a LOS of 240 days (Post et al 2005) and a South African study reported a mean LOS of 73.11 days (Joseph 2011).In addition to the occurrence of pressure sores and being unemployed, another factor that could also have influenced the LOS could be a lack of discharge settings designed for care of disabled persons including those with a spinal cord injury.In Rwanda the only discharged settings available for those who could not be discharged home are mainly for the elderly and are owned by nongovernmental and faith based organizations.A lack of discharge settings may impact the ability of indi viduals with SCI adequate care to live in the community, (Anzai et al. 2006).Data relating to discharge planning was however not collected in the current study.Another limitation of the study was that only a limited number of per sonal factors that could have impacted on LOS were investigated in this study.Further studies could determine and explore other factors such as discharge planning or availability of a discharge setting which could impact on LOS.

CONCluSION
The findings of the present study revealed that employment status, and the occurrence of pressure sores are the major factors affecting length of hos pital stay among individuals with spi nal cord injuries at Kanombe Mili tary Hospital.It seems therefore that lack of preventive measures for pressure sores, including a lack of nursing staff could have contributed to the high number of patients with pressure influencing LOS.
There is a strong need for all members of the rehabilitation team to ensure the implementation of effective measures to prevent the development of pres sure sores, in patients with spinal cord injuries in the study settings.As part of this rehabilitation team physiotherapists should ensure that these preventative measures are implemented.

Figure 1 :
Figure 1: Length of hospital stay among the sample.