chaNGes iN actiVity liMitatiONs aND PReDictORs OF FUNctiONal OUtcOMe OF PatieNts With sPiNal cORD iNJURy FOllOWiNG iN-PatieNt RehaBilitatiON

Correspondence Author: Conran Joseph University of the Western Cape Physiotherapy Department Private Bag x17 Bellville 7530 Email: conran.joseph@gmail.com AbstrACt: The purpose of this study was to investigate the changes in the activity limitations of patients following in-patient rehabilitation and the factors influencing functional ability as measured by the Spinal Cord independence measure iii (SCim iii). A longitudinal study design was utilised to study the change in functional abilities of patients with spinal cord injury between admission and discharge. A convenient sampling strategy was employed, in which every consecutive patient admitted to the rehabilitation centre within a three month period was eligible for the study. Demographic-, medical, and process of rehabilitation data were collected and collated from the patients’ medical records using a data gathering sheet that was validated and tested for reliability. functional abilities were measured by the SCim iii. Seventy-six patients met the inclusion criteria, consisting of 58 paraplegics and 18 tetraplegics. The mean age of this cohort was 34.14 years. A significant difference (p<0.001) in functional ability was detected for the total sample, with only12.5% of patients independent in walking ability and 28.12% in stair management. four (4) factors were found to be predictors of functional outcomes on bivariate analysis, but when considered together in a multiple regression model, only functional status on admission remained correlated to functional outcomes. Conclusion and implication for practice: Significant improvement in functional abilities of persons with spinal cord injury following in-patient rehabilitation was observed. however, mobility and stair-management limitations were the most prevalent at discharge. lastly, a lower functional status should be better targeted to optimise functional ability in the future. future research should be directed towards illuminating whether personal factors or rehabilitation inefficiencies are responsible for the limitations observed at discharge.

neurological impairment after acute SCI (Anderberg et al, 2007).On the contrary, there is high level evidence indicating that rehabilitation strategies could result in optimal functional outcomes (Foy et al, 2011).Despite these advancements centred towards longevity, the function ing of persons with spinal cord injury is still significantly lower than that of the rest of the population (Post et al, 1995).
Rehabilitation has become an integral part of the health care needs (Stucki et al, 2002) of persons living with spinal cord injury, with its core focus on reducing the negative consequences of the injur ing event (Staines et al, 2009).From iNtRODUctiON Spinal cord injury (SCI) is a condition that has the potential to result in severe impairments, difficulties to perform func tional tasks and a compromise in quality of life (De Vivo et al, 1999).Prior to the 1940's, physicians perceived a spinal cord injury as "...an ailment not to be treated" (Hughes, 1988).Since the First World War, there has been a dra matic improvement in the management of patients with SCI in both the medi cal and rehabilitation spheres.However, to date, there is still no good scientific evidence for treatments conducted on patients showing an improvement in an operational perspective, the goal of rehabilitation is to establish effec tive and efficient practices meeting the needs of the service user and provider (Kaplan, 2007).One of the means of determining the efficacy of rehabilita tion is by investigating the outcomes of patients, secondly by determining the factors influencing outcome and finally by aligning the process of rehabilita tion (Kaplan, 2007;Hoenig et al, 1999) to address the unique human charac teristics that were altered by the spinal cord injury.
With the acceptance of the Inter national Classification of Functioning, Disability and Health (ICF) as the pre ferred conceptual model of describ ing and understanding disability, the purpose of rehabilitation has been centred towards optimising function in all dimensions of human functioning, within the categories of impairments, activity limitations and participation restrictions (Stucki et al, 2002, WHO, 2001).Within this framework, the term "activity" is defined as the difficulty an individual encounter with the execu tion of a task or action (WHO, 2001).A study by Dahlberg and colleagues on a cohort of patients with spinal cord injury in Helsinki found that the prevalence of activity limitations was greatest for activities related to stair management and bathing (Dahlberg et al, 2003).Literature found that pre valence data on the activity limitations of persons with SCI as essential evidence to authorities, and may assist in the planning and coor dination of rehabilitation to compliment the daytoday functional needs of the patient (Dahlberg et al, 2003).
The investigation of activity limita tions could be seen as an iterative pro cess, which could lead to the adapta tion of the rehabilitation experience for specific cohorts of patients.Since an activity occurs in a context, one can infer that the execution of similar complexity of tasks within a social environment might present considerable difficulty to the patient because very few reha bilitation strategies take into account the unique context of the patient.
Kaplan considered the constituents of achieving the desired level of function ing and found that the identification of factors that may predict outcome is a critical aspect that should be determined and incorporated in appropriate health services or processes (Kaplan, 2007).Lite rature concerning the factors influ encing outcome in patients with spinal cord injury highlights factors such as

Research setting
The study was conducted at a specia lised rehabilitation facility in the Western Cape, South Africa.The centre admits patients with various neurologicaland surgical health conditions, however this study population was limited to individuals with spinal cord injuries.Team members at this facility function within the interdisciplinary approach.One feature of this approach is the regular (weekly) discussions regarding the patient's progress and the collective and collaborative goal setting of indi viduals.In addition to the structural aspect of interdisciplinary rehabilitation at the centre, the use of the rehabilita tion outcome levels (ROL's) has been regarded as standard practise and as a predictive measure of functioning along the continuum of the health condition.In short, ROL's are groupings (from level 0 to level 5) of patient problems that occur because of the consequences of the health condition along a conti nuum, i.e. from the acute through to the chronic phase.

Population and sampling
A sample of convenience was used to select the participants.Every consecu tive patient meeting the inclusion crite ria for a three month period was eligible for participation.Based on retro spective data (from the year 2010), the centre admitted more or less 320 patients with spinal cord injuries for one calendar year (personal communication, Hendry, JA).Therefore, an estimation of the desired sample size was considered for a three month period which started in November 2010 to February 2011.Eligible participants were those pre senting with a spinal cord injury, aged 18 years and above, who consented in writing to participate and were profi cient in English, Afrikaans or Isixhosa.Exclusion criteria included patients pre senting with a mental disability, which was confirmed with the inability to per form a three step command test prior to the commencement of data collection.The threemonth timeframe allowed for the desired sample size to be reached in order to measure the six predetermined factors influ encing functional indepen dence of patients with spinal cord injury age, gender, severity and aetiology of injury to be predictive of functional recovery (Scivoletto et al, 2003).Apart from personal factors influencing out come, studies reported the use of certain protocols, service and process issues to account for much of the variance in functional outcome (Foy et al, 2011).Van Hedel and colleagues compared rehabilitation and the outcomes to a "black box", explaining that very little is known about the rehabilitation pro cesses responsible for the changes in outcome (Van Hedel et al, 2012).This finding highlights the need for evalua tive studies that allows us to determine factors responsible for much variance of the primary outcomes often used in rehabilitation.
Since the ICF framework provides the baseline against which pertinent out comes should be measured, and takes into account the role of contextual factors of an individual, it is imperative to have insight into the factors pre dictive of functional outcome of indivi duals between different rehabilitation settings and patients from distinct geographical areas.In addition, the high prevalence of limitations associated with certain functional tasks will highlight the need to investigate the process and content of rehabilitation currently operational in managing those limitations.This study sets out to determine the changes in functional activity between admission and discharge, identify activity limita tions at discharge and investigate the factors influencing functional out comes of patients with a spinal cord injury following specialised inpatient reha bilitation.

Research design
The study design was longitudinal and observational in nature, in which indi viduals with SCI were followed for data collection purposes for the duration of the inpatient rehabilitation, with data collected on admission and discharge.A fundamental principle of the obser vational approach is that it does not allow for the rehabilitation process to be influenced in any way (Smith and Smith, 2003).at discharge.The literature recommends that for every independent variable entered at least ten respondents should be included.Thus, the three month period allowed for a minimum of 60 par ticipants to be included (Munro, 2001).

Data gathering sheet
The demographic, medical data as well as the information relating to the pro cess of rehabilitation and rehabilita tion outcome levels were collated from the medical records of the patients at discharge.The items for the demo graphic and medical profile included; age, gender, marital status, diagnosis, classification of spinal cord injury.Certain items were also related to impairments following the spinal cord injury, thus muscle weakness and blad der and bowel function were captured as either "impaired", "not impaired" or "undocumented".The documentation of the impairments was only retrieved by the doctors, physiotherapist or occu pational therapists assessment notes on admission.Also, the variable pertain ing to educational level was recorded as either primary, secondaryor tertiary education.The items related to the pro cess of rehabilitation and rehabilitation outcome levels were limited to the length of hospital stay (LOHS), discharge des tination, health professionals involved in the management of the patients, partici pation in the learntoswim programme and outcome levels achieved on admis sion and at discharge.
The data gathering sheet was validated for content by two experts in the field of rehabilitation and was also found to be reliable (interrater) with ICC scores ranging between 0.741.00 for the com pleted sheet.Thus, the data gathering sheet presented with excellent reliability in the instance where documents were reviewed by two independent raters.

Measure for functional independence and translations
The Spinal Cord Independence Measure (SCIM) III was used to assess functional ability in this cohort of patients with spinal cord injury.This valid and reliable tool was chosen as it is the first spinal cord injury specific measure developed for measuring functional outcomes in this particular health condition (Catz and Itzkovich, 2007).Also, all items of this measure can be mapped to the domains of the ICF, which is the concep tual model used to define activity and that was used to underpin the research findings.The SCIM III consists of 19 items, which is divided into four areas namely, selfcare, respiration, sphincter management and mobility.The out come measure is scored on an ordinal scale, with the cumulative score ranging between 0100.Since this measure was developed with the European context in mind, it is deemed imperative to test the psychometric properties within a dif ferent context and that of the translated versions.The languages spoken by peo ple in Cape Town are English, Afrikaans and IsiXhosa, thus the SCIM III was translated into Afrikaans and IsiXhosa.All the questionnaires were forward translated from English to Afrikaans and English to IsiXhosa respectively by one translator, thereafter the translated questionnaires were backward transla tion into the original language (English) by a another translator.Both translators were independent to the research pro ject and not familiar with the medical or rehabilitation field.The SCIM III was scrutinised by the researcher and study supervisor for correctness and whether the content of each item remained the same irrespective of the translation by mapping the main construct of each item back to domains of the ICF.The reason for this process is that it is often found than when an outcome measure is translated into a different language, the translated language may express and interpret the items differently than the original version (ChangHoon, Dong Jae, SeKang, DongJun, HwanMo & HeuiJeon, 2006;Mkoka, Vaughan, Wylie, Yelland & Jelsma, 2003).There fore, based on consensus among the researcher and both translators the fol lowing changes were made to the SCIM III.The IsiXhosa translation of the SCIM III presented with the following changes.The word "oksijini" meaning oxygen from the natural environment/ atmosphere was changed to "umoya omncedisaya" which refers to artificial oxygen produced by a mechanical ven tilator.This change was imperative as this would have influenced the results of respiratory function among spinal cord injuries as the aforementioned words have contrasting meanings.The word "oksijini" has an active component to breathing, whereas the desired word "umoya omncedisaya" has a passive or assisted component to respiration.The IsiXhosa word describing mobility was changed from "iintshukumo", meaning shaking, to "ukuhla usenyuka", which refers to walking as a higher functional task.Thus, the word was originally used to describe a patient's ability to walk as a means of mobilisation.The measure can be administered in different formats while maintaining good to excellent reli ability, but for the purpose of this study the administration was via interview (the lead author and a trained research assistant who is proficient in English and IsiXhosa) which was self reported (Itzkovick et al, 2003).A recent study reported on the high correlation between the SCIM III and the SCIM self reported version (Fekete et al, 2013).

Procedure and ethical statement
All patients meeting the inclusion cri teria were invited to participate in the study.After obtaining written informed consent from the patients, the medical records of the patients were perused, and a time slot was decided on for the completion of the functional outcome measure via interview.The data gather ing sheet was completed on admission with the functional assessment done on admission and discharge via interview.All patients consenting to participate in the study were interviewed within two days of admission (the first two days are usually focused on the evaluation of the patient, with little to no rehabilita tion input) and within the last two days prior to their discharge.All interviews lasted between 3045 minutes.The methodology for this study had been ethically approved by the University of the Western Cape's Senate Research Committee (10/3/24), and permission was granted by the Department of Health and the Deputy Director at the Western Cape Rehabilitation Centre.

statistical methods
All data elements were coded and cap tured twice in Windows Excel 2007, and imported to SAS for analysis.The Shapiro Wilk W test was used to deter mine the normality of data elements.With reference to the measurement of outcome on admission and discharge, the outcome variable which was the difference in score was found to be normally distributed.Thus, the paired ttest was used to determine the signifi cant difference in functional outcome of patients between admission and dis charge, with the use of the independent ttest for the calculated mean differences between subclassifications of the spinal cord injured population.A 5% significant level was used for all inferential statis tics, which translates to an alpha level of 0.5 (Domholdt, 2000).Data related to the functional ability of patients with spinal cord injury were collected accord ing to completeness of injury and each level of lesion.However, due to the small sample sizes (<10) for each level, inferential statistics were not advised, due to limited power of analysis.
In response to one of the aims of the study which was to determine the fac tors influencing functional outcome, both linear and multiple regression models were considered to determine the relationships between independent varia bles and the dependent variable (functional outcome score as measured by the SCIM III).Based on the study of mean outcome scores for different categories, data were found to be abnor mality distributed, thus nonparametric statistics were used.For dichotomous predictors (e.g.gender) a Wilcoxon Rank sum test was used to compare the mean outcome scores between dichotomous categories, and the Spearman's rank order correlation coefficient was used for continuous independent variables such as length of hospital stay, age, functional outcome score on admission (0100).A multiple regression model was fitted for four individual indepen dent variables which were found to be associated with the functional outcome score of patients with spinal cord injury as measured by the SCIM III.The four independent variables were considered together and the final model was esta blished through the backward elimina tion process.

Recruitment of participants
Patients with spinal cord injury accounted for 40% of the total admis sions to the WCRC for the period of November 01, 2010 to January 31, 2011.Of the 93 patients with spinal cord injury, nine patients were under the age of 18 years old, two declined consent, five subjects presented with missing data and one presented with a mental illness as comorbidities, which was considered as part of the exclusion criteria.Thus, the sample meeting the inclusion criteria was 76 patients with spinal cord injury.For the study of difference in functional outcome between admission and dis charge, the sample size was reduced to 64.The reasons include the unexpected discharge of patients outside the Cape Metropole area (5 patients), suspension of patients following their weekendpass to their residence (2 patients), death while hospitalised (2 patients) and the referral of patients to tertiary hospitals due to secondary complications.

Profile of the participants
The demographic data are presented in Table 1.The mean age of the patients was 34.14 years (SD: 12.42), with more paraplegic patients (76.3%) follow ing injury to the spinal cord.The age of patients ranged from as young as 18 to 81 years of age.Most of the patients were single at the time of injury.

Process of rehabilitation
All patients with spinal cord injury were managed by the nurse and physiothera pist, with most of them also managed by the doctor, occupational therapist and social worker.The length of hospital stay in days was longer for the cohort with tetraplegia (85.6), with the majority of patients discharged to their longterm residence.Furthermore, less than 20% of the sample was referred for follow up medical intervention at either the institution's outpatient department or another health care facility.Several of the patients (40.8%) were admitted with a health status corresponding to outcome level two, and 69.7% of patients were discharge with an outcome level greater than three (Table 2).

Functional outcomes of patients according to the spinal cord independence Measure iii
Table 3 illustrates the statistical dif ference in functional outcome between admission and discharge, and provides the level of significance for the total sample as well as the subclassifications of SCI.The sum score on the SCIM III at discharge, for both the total sample and subclassifications, was significantly larger than that on admission as deter mined by the paired ttest.The sig nificant difference between the means scores of those with paraplegia and tetra plegia indicated that the mean score of the paraplegic population, on both admission and discharge, was signifi cantly higher than the tetraplegic group with p=0.023 and p=0.007 respectively.

independent execution of tasks as derived from the sciM iii on admission and at discharge
Figure 1 shows the degree of indepen dent execution of tasks on admission was greatest for respiration, feeding and grooming with 98.7%, 79.0% and 76.3% of patients fully functional with the respectively tasks.The tasks in which most patients were still dependent on and needed assistance with, were related to mobility (indoors, along moderate distances and mobility outdoors), stair management and transfer from floor to wheelchair.The greatest increases in the inde pen dent execution of tasks between admission and discharge were found for bathing of the lower limbs, transfer from the wheelchair to the car, and toileting with 43.7%, 43.3% and 43.0% of addi tional patients who were not indepen dent with the execution of the respective tasks on admission.

Factors predictive of outcome for patients with spinal cord injuries
The SCIM III was used to measure func tional abilities in the cohort of patients with spinal cord injuries.This measure was considered as an ordinal scale with the score ranging between 0100 in increments of 5 points.The list of independent variables (which was both continuous and categorical) that was considered for the bivariate analysis is summarised below in table 4.
The results indicate that higher acti vity scores on admission, younger age, and shorter length of hospital stay were found to be correlated with better func tional abilities, as measured by the SCIM III, at discharge.The inverse rela tionship between age and length of hos pital stay and the functional outcome score at discharge is evident in the Spearman's correlations of 0.30 and 0.28 respec tively.The only categorical variable that was found to be correlated with func tional ability were the absence of partici pation in the learnto swim programme offered at the centre, as this was con firmed by a p=0.013.The relationship was strong between the acti vity levels on admission as measured by the SCIM III compared to the activity level at dis charge.This relationship was confirmed by Spearman's r=0.66 and a p<0.001.

Multiple predictor model considered for spinal cord injury
Since there were two variables that showed a relatively strong relationship to outcome level scores (as measured by the SCIM III) and two other variables showing a slightly weaker relationship, but yet significant, these variables were considered together in a multiple regres sion model.(See Table 5).
During the backward elimination process age, participation in the learn toswim programme and length of hos pital stay were insignificant predictors when considered with only the outcome score on admission.With regards to the parameter estimate of the outcome score on admission, the analysis indicates that for every one point (as achieved on the SCIM III) on admission, the predicted value of the outcome score at discharge (as measured by the SCIM III) will have an additional 0.57 points.

DiscUssiON
The aim of this study was twofold; firstly to determine the activity limita tions and secondly to identify the predic tors of functional outcome as measured by the SCIM III.The mean intake SCIM III score for this cohort of patients was 48.03, which clearly shows that partici pants in the present study had a much higher level of functioning than patients with SCI in Europe, with a mean intake SCIM II score of 32.00 (Wirth et al, 2008).The latter reported study demon strated change in functional status at three and six months with median SCIM II scores of 60 and 71 respec tively.It is important to note that the mean LOHS of patients with spinal cord injury in the current study was 73.11, which loosely converts to more or less two and a half months of hospitalisation, but with a mean SCIM III score of 67.07 (SD: 23.0), which was higher than the score obtained after three months in the study conducted by Wirth et al (Wirth et al, 2008).A possible explanation for this difference in functional status between the current study and the study by Wirth and colleagues could be that the sample in the latter study comprised of more complete tetraplegics (42% of sample), whereas the current study included more paraplegics (76.3%), or the criteria for admission to the centre could have accounted for this difference in which only persons with no complications or secondary complications are admitted and those who could benefit from in patient rehabilitation.
The highest levels of independence were for the items respiration, feed ing and grooming.These findings are consistent with the results from a study conducted in eight rehabilitation centres in the Netherlands (Post et al, 2005), except for the item relating to respira tion.The reason for this being that the latter study utilised the FIM, which omits items related to respiratory func tion.Looking at individual items, the majority of the patients with spinal cord injury in the current study experienced high levels of limitation with mobi lity (walking) and stair management at discharge, nevertheless the difference was statistically significant between the points of data collection.Similar findings were found elsewhere (Wirth et al, 2008;Chan and Chan, 2005).Optimal outcome levels have not yet been achieved at the time of discharge from rehabilitation for this cohort of patients.The tacit knowledge as derived from the items of the standardised outcome measure and the global rating score of the rehabilitation outcome levels sug gests that there are limitations that could be addressed.However, the authors are in agreement that a qualitative inquire would yield considerably more trust worthy data of the nature of the par ticipants' problems and the barriers to their achievement of certain activities or engagement in certain life situations.
The high level of disability associated with mobility (specifically relating to walking ability) is slightly ambiguous as the nature and extent of the spinal cord injury were not taken into account, irrespective of the independent execu tion of mobility tasks using a wheelchair.Despite the recognition of the SCIM III as the gold standard measure in the evalu ation of the functional abi lity of persons with SCI, the measure lacks responsive ness.For example, if a patient is capable of being mobile in a wheelchair, without physical assistance, the outcome meas ure will continue to indicate an inferior score compared to someone walking with assistance and personal support.Despite this limitation, the need for follow up rehabilitation post discharge is implied since studies found significant improve ment in mobi lity, particularly within the mobility construct, for up to six months after injury (Wirth et al, 2008).Since patient with spinal cord injuries are inherently limited in mobility tasks, the author and other published authors (Ullrich et al, 2012) are concerned about current measuring tools, such as the Functional Independence Measure and SCIM III that do not consider the level of lesion as a legitimate level of func tioning with predefined capabilities for that specific individual.Therefore, floor and ceiling effects of outcome measures are almost always present.The recom mendation of a recent study by Ullricht and colleagues advocate for the inclusion of patients with spinal cord injuries in the decisionmaking process of which con structs of functioning should be included for the evaluation of the holistic impact of such an injury on the individual (Ullrich et al 2012).
With the focus on predictors of functional outcome in this cohort of patients, the current study found age to be inversely related to functional out come.Case controlled studies conducted by Scivoletto et al (2003) and Seel et al (2001) found that independence of daily living measures were significantly greater (p<0.001) for the younger age group and the young to middle age group respectively.The achievement of better functional outcomes in the younger age group could be due to the traumatic nature of the lesions, which have been associated with a higher probability of sustaining complete motor lesions of the thoracic and lumbar levels, whereas nontraumatic lesions are more com mon in older patients and have higher frequencies of incomplete tetraplegia (Scivoletto et al, 2003).Apart from age, the current study and numerous pu blished reports found length of hos pital stay to be a predictor of functional outcome at discharge (Sipski et al, 2004;Scivoletto et al, 2003).The current study found a significant correlation between LOHS and functional outcome at dis charge (spearman's rho0.28;p=0.035), which in this instance could be the result of the sample consisting of more para plegics (76.31%) than tetraplegics (Post et al, 2005).Although the mean length of hospital stay in the current study between those with paraplegia (68.53;SD 50.5) and tetraplegia (85.62;SD 40.6) was statistically insignificant, the difference of 17 days has clinical meaning and rele vance as cost of healthcare is usually estimated based on the length of hospital stay in days (Winslow et al, 2002).The inability to detect the significant diffe rence in LOHS between those with para plegia and tetraplegia could be due to the large standard deviation, which could be influenced by the small size that reduces the power of analysis.The shorter length of hospital stay as an indicator for better functional outcomes should be analysed with caution, as it could simply imply that those with paraplegia reach their levelspecific outcomes sooner than those with tetraplegia, due to the initial level and severity of the injury and age.
Literature highlights that by com bining landbased rehabilitation with aquatic physical therapy, better outcomes may be achieved in recovery of function in adults with acquired neurological impairments (Degano and Geigle, 2009).The current study findings provided evi dence of the effect of a learntoswim programme /aquatic therapy on the func tional outcomes of those patients with spinal cord injury.The bivariate analysis demonstrated that those who attended the learntoswim programme had a significantly higher functional outcome (p>0.013) at discharge than those who did not attend, whereas an insignificant difference between the mean functional outcome scores of the two independent samples (those who did attend versus those who did not attend) was found on admission.
Results from the multiple regression model, which used a backwards elimi nation procedure indicated that only functional outcome to be significant in the model.Similar results are found elsewhere (Burnett et al, 2000).Based on theoretical principles, discharge plan ning should be guided by functional out come and not health care cost (Visagie, 2012).On the basis of this investigation, with the author aware of the limitations of this model (small sample size which led to the inclusion of only certain vari ables), the functional outcomes at dis charge could be predicted.This model provides relevant health care profes sionals and patients to aid in the iden tification of factors which could impede functional outcomes and the opportunity to aggressively manage and appropri ately refer patients to other relevant pro fessionals.Since the functional outcome on admission accounted for almost half 46% of the variance in the overall func tional score at discharge, it is recom mended that rehabilitation professionals administer the SCIM III on admission, which is an easy tool to use, in order to set realistic goals and to decide on the appropriate rehabilitation services and processes for the patient.
In conclusion, this study clearly high lights the changes in functional abili ties and activity limitations in patients with spinal cord injury, following spe cialised inpatient rehabilitation.Based on the outcomes, limitations in mobi lity and stair management are still most prevalent at discharge.Since activity limitations influence participation, given that participation restrictions are dif ficult to be evaluated and addressed outside of the patient's social milieu, it is essential that rehabilitation services should be available, not necessarily on an inpatient basis, following discharge from the specialised rehabilitation unit.Furthermore, the predictors of functional outcome of this particular study should serve as a basis for targeting and appro priately manage the factors that could negatively impact the functional out come of patients with spinal cord injury.

liMitatiONs OF the stUDy
This study did not determine the com pleteness of the lesion and failed to capture the neurological classification according to ASIA.Consequently, com parison of data will be negatively influ enced.Furthermore, the exclusion of patients under the age of 18 and those with comorbidities could have influ enced the normality of the data.

acKNOWleDGeMeNt
The authors would like to thank South Africa Netherlands research Pro gramme on Alternatives in Develop ment (SANPAD) for financial support and for the patients for their partici pation in this study.ReFeReNcesAnderberg L, Aldskogius H, Holtz A 2007 Spinal cord injuryscientific challenges for the unknown future.Ups J Med Sci 112: 259288 Burnett DM, KolakowskyHayner SA, Gourley EV, Cifu DX 2000 Spinal cord injury "Outliers": An analysis of etiology, outcomes, and length of stay.J neurotrauma 17(9): 765772 Catz A, Itzkovich M 2007 Spinal Cord Independence Measure: Comprehensive ability rating scale for the spinal cord lesion patient.J Rehabil Res Dev 44(1): 6568 Chan SC, Chan AP 2005 Rehabilitation out comes following traumatic spinal cord injury in a ter tiary spinal cord injury centre: a comparison with an international standard.Spinal Cord; 43(8): 489498 Dahlberg A, Kotila M, Kautiaien H, Alaranta H 2003 Functional independence in persons with spinal cord injury in Helsinki.J Rehabil Med 35: 217220 Degano AC, Geigle PR 2009 Use of aquatic phy sical therapy in the treatment of balance and gait impairments following traumatic brain injury: A case report.J Aquatic Phys Ther 17: 1621 DeVivo MJ, Krause JS, Lammertse DP 1999 Recent trends in mortality and causes of death among persons with spinal cord injury.Arch Phys Med Rehabil 80: 1411-1419 Domholdt E 2000 Physical Therapy Research: Principles and Application (2 nd ed.) WB Saunders Publishers: Philadelphia Fekete C, EriksHoogland I, Baumberger M, Catz A, Itzkovich M, Luthi H, Post MWM, von Elm E, Wyss A, Brinkhof MWG 2013 Development and validation of a selfreport version of the Spinal Cord Independence Measure (SCIM III).Spinal Cord 51: 4047 Foy T, Perritt G, Thimmaiah D, Heisler L, Offutt J, Cantoni K, Hseih CH, Gassaway J, Ozelie R, Backus D 2011 Occupational therapy treatment time during inpatient spinal cord injury rehabi litation.J Spinal Cord Med 34(2): 162175 Hoenig H, Horner RD, Duncan PW, Clipp E, Hamilton B 1999 New horizons in stroke rehabili tation research.J Rehabil Res Dev 36(1): 1931