LENGTH OF HOSPITAL STAY OF ORTHOPAEDIC INPATIENTS IN A TERTIARY CARE TEACHING HOSPITAL IN NEPAL

Background : Length of Hospital Stay (LOHS) can have important effects on the cost of treatment and patient outcomes. The aim of this study was to determine the length of hospital stay among orthopaedic inpatients and assess its association with different socio- demographic and clinical factors. Methods: A retrospective study was conducted at College of Medical Sciences and Teaching Hospital, Nepal, wherein, clinical records of patients admitted and treated as inpatients between January and December 2019 were retrieved. Demographic data, diagnosis, treatment details, LOHS, co-morbidities, treatment modality and mode of payment were documented and data was analyzed using SPSS software 16.0. Median was calculated for skewed continuous data and frequency analysis was done for categorical variables. P value <0.05 was considered statistically significant. Results: In a total of 1248 patients with mean age of 33.8±18.7 years, 72.4 % (n=904) were male. Most of them (34.9%, n=435) were students and majority (62.5%, n=780) were from outside Chitwan. The median LOHS was 5.0 (3.0–10.0) days. It was significantly greater in cases from outside Chitwan, those with trauma, infection, associated injury, and complications (p<0.05). Moreover, there was significant difference in LOHS among various age quintiles, occupations, fracture types, modes of payment and treatment modalities (p<0.05). However, LOHS did not differ significantly between patients with or without co-morbidity and gender (p>0.05). Conclusions: This study identified that the LOHS was significantly associated with various clinico demographic factors except gender and co-morbidity. More studies can be conducted to assess the relationships further.


INTRODUCTION
Length of hospital stay (LOHS) can be defined as the total length of time a patient stays in the hospital for the purpose of treatment, measured as the interval (usually in days) from admission to discharge. 1 Burden of disease in the Orthopaedic discipline, specifically complex polytrauma related morbidities, is on the rise. 2 In this regard, LOHS can have important effects on the cost of treatment, rate of infection, risk of bed sore and other co-morbidities, specifically amongst the orthopaedic inpatients. 3 Various studies conducted far and wide have reported fairly varied ranges of LOHS in these patients. In two separate studies conducted in Canada and the USA, average LOHS as high as 9.4 days and as low as 4.3 days were reported with significant associations with age, gender, co-morbidities and insurance status. 4,5 In two other studies conducted in Iranian population, mean LOHS of 5.4±6.1 days 6 and 6.8±8 days 7 were reported. In the context of Nepal, Mishra et al. 8 observed that the average LOHS among orthopaedic inpatients was 10.5 days and was associated with age, gender, mode of payment and type of cases (traumatic/non traumatic).
There is paucity of literature regarding LOHS and associated factors to influence the LOHS in our context. So, this study was undertaken to find out the LOHS among the orthopaedic inpatients in a tertiary care teaching hospital and variables affecting it.

METHODS
This was a retrospective study conducted in the department of orthopaedic surgery at College of Medical Sciences and Teaching Hospital (COMS-TH), Bharatpur, Chitwan, Nepal. After obtaining ethical clearance from the Institutional Review Committee of COMS-TH (COMSTH-IRC) (Ref No: 2020-041), clinical records of the patients were extracted from the medical record section of the hospital.
Various socio-demographic and clinical information of the patients who were admitted and treated as inpatients in the hospital from January to December 2019 were retrieved. Patients who left the hospital against medical advice (LAMA), those with incomplete data, and those who expired during treatment in the hospital were excluded. A total record of 1248 patients were selected for the final analysis. Data was obtained from the details entered into the admission and treatment charts of the inpatients. The different sociodemographic variables collected included age, gender, occupation, address, length of hospital stay. Similarly, the clinical variables were type of cases (traumatic/non-traumatic), infective cases, type of fracture (open/closed/ combination of open and close), associated injuries, complications, comorbidities, treatment modality (conservative/operative/both) and mode of payment (self/third party/health insurance). Age of the patients was discretized into five quintiles and LOHS into two categories (LOHS ≤7 days /> 7 days).
The collected data were first entered into the Microsoft Excel (Microsoft Office 2010). After preliminary cleaning, data was entered into SPSS (Statistical Package for Social Sciences), version 16.0 software for final data analysis. The various categorical variables were described as frequency and percentage using appropriate tables. Next, the distribution of the primary variable of interest, LOHS was tested for normality. As the distribution was found to be skewed, it was described using median (inter-quartile interval). To test the association between LOHS (as a continuous variable) and various categorical variables, non-parametric tests like Mann-Whitney and Kruskal-Wallis H tests were utilized. Similarly, Chi-squared test was used to test the association between two categories of LOHS and other categorical variables. Statistical significance was defined as p< 0.05 at 95% confidence interval (CI).

RESULTS
In the present study, out of the total patients (n=1248) , males constituted the majority, i.e. 72.4%, (n=904). The mean age of the patients was 33.8 years ± 18.7 years, the median age was 30.0 years (range 9 months-95 years). Regarding occupation, most of the patients were students (34.9%, n=435), and homemakers were the least common (15.3%, n=191). Likewise, majority of the patients were from outside Chitwan (62.5%, n=780).  The median length of hospital stay was 5.0 days (range: 1.0 -104.0 days). Table 3

*Kruskal-Wallis H Test
The LOHS was significantly greater in cases with trauma (p=0.001), with infection (p<0.001), with presence of associated injury (P<0.001), and with complication (p<0.001) than those without the above conditions. Moreover, there was significant difference in LOHS among various fracture types, mode of payment and treatment modality (p<0.001). However, LOHS did not differ significantly between patients with or without co-morbidity (p=0.763).  The LOHS of the patients were categorized into (a) ≤ 7 days and (b) >7 days. Proportions of patients with LOHS greater than 7 days did not differ significantly across the gender, and presence/ absence of comorbidity (p>0.05). LOHS was significantly different across the various age quintiles (p=0.003). Across the various occupations, the proportion was the greatest in miscellaneous groups followed by farmer, homemaker and students. The overall difference was statistically significant (p=0.018). Similarly, the proportion was statistically greater in patients from outside Chitwan (p=0.002), patients with traumatic injuries (p<0.001), infection (p<0.001), associated injuries (p<0.001), and complications (p<0.001). Similarly, the proportion of patients staying for more than 7 days in the hospital was significantly different across type of fracture, mode of payment and treatment mo-dality (p<0.001). However, there was no significant association between co-morbidity and LOHS (p>0.05).  10 also reported proportionately increased LOHS with increase in age of the patient. This could be explained by tendency to early discharge in younger age group patient.
The different occupation groups in our study had significantly different median LOHS (p=0.001), students and homemakers having the shortest LOHS. Similarly, the prevalence of LOHS>7days also differed significantly across these groups (p=0.018). The study done by Haghparast-Bidgoli et al 7 observed that LOHS among patients with road traffic injuries was longer in manual workers and farmers compared to other occupation groups (p<0.001). Similarly, the study done by Khosravizadeh et al 6 found that the mean LOHS was longest for retired patients and shortest for students (p<0.01).
Patients from outside Chitwan had significantly higher median LOHS and also the higher prevalence of LOHS>7days than those residing within Chitwan (p<0.05). This could be due to tendency for asking early discharge for the patients residing near the hospital. Wu et al 11 observed that the LOHS was significantly associated with geographic location of the hospital and was greater in suburban compared to urban patients (p<0.05). Sukumar et al, 12 on the other hand, observed that people staying in rural area with fall related injury had shorter median LOHS compared to people staying in urban area (P<0.05).
Patients presenting with trauma had significantly greater median LOHS (25% more than nontrauma) as well as the prevalence of LOHS>7days, than those without trauma (p<0.05). Comparable finding was also observed by Mishra et al 8 who reported that the average LOHS was longer in patients with trauma (11 days) compared to those without trauma (8 days).
Patients with infection also had significantly higher median LOHS and prevalence of LOHS>7days than those without infection (p<0.05). This finding was consistent with many other studies. Glance et al 13  Limitations of the current study was retrospective chart review nature of the study which might limit the extraction of every variables such as ICU admissions, malnutrition, depression, etc as they might influence the LOHS. Secondly, there is lack of generalizability of the result since it is a single institution-based study. The strength of this study is its sufficient sample size of one year duration thus enabling the samples with all possible spectrums of orthopaedic conditions across all seasons of the year.