Factors predicting burn unit length-of-stay
Fatores preditivos da permanência em uma Unidade de Queimados

INTRODUCTION

Due to a generalized improvement in healthcare, the last 30 years presented us with a significant reduction in the mortality rate of burn patients¹,². Several factors are responsible for this change such as a better understanding of the physiopathology of severe burns, the widespread development of critical care units, and new patient-tailored therapeutic strategies, namely, aggressive fluid resuscitation, rigorous infection control and early surgical intervention with thorough debridement and skin grafting to surpass the loss of assessment power of burn-related mortality.

Therefore, to evaluate the quality and efficiency of clinical care, it is important to assess other outcomes such as length of stay³ whose increase is irrefutably associated with adverse consequences. A longer stay in a burn unit or intensive care unit has been associated with a higher physical (and psychological morbidity, a delayed return to work, decreased productivity, lower health-related quality of life, and a higher and a higher incidence of psychopathological symptoms⁴,⁵.

Burn patients are particularly susceptible to infection. A prolonged length of stay increases the risk of nosocomial infection, which in turn increases the length of stay in a burn unit by an average of 18 days. Furthermore, an increased length of stay has been proven to be directly related to the development of antibiotic resistance. The most common infections in this context are caused by aggressive microorganisms such as Pseudomonas aeruginosa and Methicillin-Resistant Staphylococcus aureus. Infection by this microorganism is associated with increased all-cause mortality and morbidity.

At last, an increased length of stay is associated with an increased social and economic cost. In Portugal, the average cost of a stay in a Burns Unit in 2013 was 8032 euros⁶,⁷. The following factors have been associated with a more prolonged stay: age⁸,⁹, male sex⁹, percentage of burnt surface area⁸,⁹, depth of burn¹⁰, presence of airway injury⁹, comorbidities or associated traumatic injury¹¹, need for a surgical procedure⁸,¹⁰ and the presence of infection or sepsis¹⁰.

In this retrospective study, the authors aim to establish what factors are relevant to the length of stay in a Burn Unit in Portugal, to identify specific measures and interventions that might allow the reduction of hospitalization time and therefore the morbidity and mortality associated with it.

MATERIALS AND METHODS

In this retrospective study, all patients admitted to the Burnt Patient Special Care Unit in the Hospital Universitário Lisboa Central between the 1^st of January 2011 and the 31^st of December 2020 were assessed. A total of 745 patients was collected. Patients were excluded from the study due to early discharge to another unit for medical or logistic reasons¹⁰; discharge against doctor orders³ or admission with a non-burn condition (toxic epidermal necrolysis)².

The data from the resulting 711 patients was collected regarding age, sex, previous medical conditions, date of admission and discharge of the burn unit, mechanism of burn, agent of burn, area of burn, depth of burn, associated traumatic injuries, need for ventilation, need for fasciotomy or escharotomy, airway injury, need for and timing to surgical intervention, infection and site of infection, laboratory parameters and clinical outcome. Statistical analysis was performed using SPSS for Windows. A p-value <0,05 was considered significant. (Table 1)

Table 1 - Burn unit population.

RESULTS

Duration of Stay

Duration of stay in the Burn unit averaged 29 days. The minimum was 1 day and the maximum was 254 days. The following factors were associated with an increased length of stay: female sex (p-value 0,048); age (p-value 0,002); and presence of co-morbidities (p-value 0,015). (Table 2).

Tabela 2 - Análise dos fatores que influenciam o Tempo de Permanência. (T – teste-t; A – ANOVA unidirecional; S – classificação de Spearman).

	Statistical Average	p value
Sex
Male	27	0,048 T
Female	31
Age		0,002 S
Age Group – years old
<20	16
21-40	25	0,007 A
41-60	30
61-80	33
81-100	31
Medical Comorbidities
Yes	31	0,015 T
No	27
Burn Mechanism
Thermal	30	0,000 A
Electrical	20
Chemical	17
Thermal Burn
Yes	30	0,000 T
No	19
Electrical Burn
Yes	20	0,001 T
No	30
Chemical Burn
Yes	17	0,048 T
No	29
Fire Injury
Yes	35	0,000 T
No	23
Hot Líquid Injury
Yes	24	0,000 T
No	31
Contact Burn Injury
Yes	23	0,167 T
No	29
Burn Injury Agent
Fire	35
Hot Liquid	24
Electric	20	0,000 A
Contact Burn	23
Chemical	17
Body Surface area		0,000 S
Body Surface area (%)
< 10%	20
11-20%	32
21-30%	45
31-40%	48
41-50%	71	0,000 A
51-60%	94
61-70%	40
71-80%	32
81-90%	61
> 90%	5
Body Surface area up to 10%
Yes	20	0,000 T
No	40
Body Surface area from 10 to 20%
Yes	24	0,000 T
No	52
Body Surface area from 20 to 30%
Yes	26	0,000 T
No	58
Body Surface area from 30 to 40%
Yes	27	0,000 T
No	64
Body Surface area from 50 to 60%
Yes	28	0,000 T
No	60
Body Surface area from 50 to 60%
Yes	29	0,012 T
No	44
Body Surface area from 60 to 70%
Yes	29	0,018 T
No	45
Body Surface area from 70 to 80%
Yes	29	0,005 T
No	54
Body Surface area from 80 to 90%
Yes	29	0,344 T
No	5
3rd Degree Burn
Yes	37	0,000 T
No	20
Need for Mechanical Ventilation
Yes	36	0,000 T
No	26
Duration of Mechanical Ventilation – n=208		0,000 S
Duration of Mechanical Ventilation > 12 days
Yes	57	0,000 T
No	28
Inhalatory Injury
Yes	43	0,000 T
No	26
Associated Trauma
Yes	34	0,125 T
No	29
Need for Escharotomy
Yes	50	0,000 T
No	25
Need for surgical intervention
Yes	36	0,000 T
No	13
Number of Surgical Interventions – n = 505		0,000 S
Timing of 1st Surgical Intervention – n=505		0,000 S
Timing of 1st Surgical Intervention – n=505
First 5 days	27
First 10 days	37	0,000 A
First 15 days	35
First 20 days	40
After 20 days	61
First 5 days
Yes	27	0,000 T
No	39
First 10 days
Yes	33	0,003 T
No	40
First 15 days
Yes	34	0,000 T
No	48
First 20 days
Yes	34	0,000 T
No	61
Documented Infection
Yes	40	0,000 T
No	20
Nosocomial Infection
Yes	46	0,000 T
No	19
Number of Documented Infections – n=328		0,000S
Number of Documented Infections – n=328
1-2 Infections	30
3-4 Infections	52	0,000 A
5-6 Infections	60
Over 6 Infections	107
Mucocutaneous Infections
Yes	41	0,000 T
No	26
Respiratory Infection
Yes	50	0,000 T
No	27
Urinary Infection
Yes	50	0,000 T
No	22
Systemic Infection
Yes	51	0,000 T
No	24
Minimum Hemoglobin Value (g/dl)		0,000 S
Anemia (Hb < 8 g/dl)
Yes	43	0,002 T
No	28
Maximum Creatinine Value (mg/dl)		0,002 S
Acute Kidney Disease (Creatinine > 1,2 mg/dl)
Yes	38	0,000 T
No	28
Minimum Protein Value (g/l)		0,000 S
Hypoproteinemia (Total Protein < 60g/l)
Yes	32	0,000 T
No	19
Minimum Albumin value (g/l)		0,000 S
Hipoalbuminemia (Albumin < 35g/l)
Yes	32	0.000 T
No	16
Threat to Life
Very Low (ABSI 2-3)	12
Moderate (ABSI 4-5)	20
Moderately Severe (ABSI 6-7)	29	0,000 A
Serious (ABSI 8-9)	42
Severe (ABSI 10-11)	73
Maximum (ABSI 12 ou Superior)	42
Mortality > 50% according to ABSI (> 10)
Yes	63	0,000 T
No	27
Baux Score		0,000S
Modified Baux Score		0,000S
Mortality > 50% according to Modified Baux Score (≥ 140)
Yes	55
No	29
Death
Yes	33	0,283 T
No	29

T – t-test; A – one-way ANOVA; S – spearman rank.

Tabela 2 - Análise dos fatores que influenciam o Tempo de Permanência. (T – teste-t; A – ANOVA unidirecional; S – classificação de Spearman).

Regarding the lesions, the following factors were associated with an increased length of stay: Thermal burns (p-value 0,000), especially if the thermal agent was fire (p-value 0,000); third-degree burns (p-value 0,000); need for mechanical ventilation (p-value 0,000); established airway injury (p-value 0,000); need for decompressive fasciotomies or escharotomies (p-value 0,000) and percentage of body area (p-value 0,000). Of importance is the comparison between early extubated patients (first 12 days) and non-early extubated, who had a significantly (p-value 0,000) longer length of stay (average of 28 vs 57 days).

The presence of associated traumatism was not significantly associated with an increased length of stay (p-value 0,125). (Table 2)

Appropriate debridement with or without grafting is the cornerstone of burn treatment. In our sample, the need for surgery was significantly associated with an increased length of stay (p-value 0,000), as was the timing of the surgery (p-value 0,000). Patients submitted to surgery in the first 5 days had an average length of stay of 27 days, however, when the surgery performed after the 20^th day, the average stay was 61 days.

The presence of infection (p-value 0,000), the number of infections (p-value 0,000), and the presence of a nosocomial microorganism (p-value 0,000) were positively correlated with the length of stay. (Table 2).

The presence of anemia (p-value 0,002); renal failure (p-value 0,000), hypoproteinemia (p-value 0,000), and hypoalbuminemia (p-value 0,000) were also positively correlated with an increased length of stay.

The classic burn prognosis indexes such as the Abbreviated Burn Severity Index and the Modified Baux Score were strongly correlated with the length of stay (p-value 0,000).

DISCUSSION

The evolution of burn care has led to an increased survival rate of burn patients which has made some classic indexes outdated in terms of prognosis. On the other hand, length of stay is an objective variable, that is easy to measure and compare and is gaining more relevance as a metric in the assessment of quality of health care in burn patients. A more prolonged length of stay in a burn unit is associated with a greater number of infections, greater morbidity, greater mortality, and costs. Therefore, it is relevant to understand which factors contribute to the increase of this metric, to access potential actions to tackle this problem.

In our study, the average length of stay was 29 days, with a range from 1 to 254 days. Compared with the current literature¹², this value is higher. However, in our unit, this data represents only the patients admitted in the Burn Care Unit, which is equivalated to an intensive care unit, and therefore most burns patients are severe and not amenable to standard infirmary care which most published studies focus on.

Similarly, to other studies, some factors associated with increased length of stay were: age, comorbidities, and burn severity-related factors, such as area of burnt skin, depth of burn, and need for decompressive escharotomy or fasciotomy. Inhalatory lesion is, in the current literature, the factor with the biggest impact on burn patients. In our cohort, the presence of airway injury with or without the need for mechanical ventilation was associated with an increase in length of stay. Similarly, early extubation was associated with a decreased length of stay.

Opposite to current literature, in our cohort female sex was associated with a statistically significant increased length of stay. This may be due to the older age of women averaged to man, leading to more burning in advanced age, where a more frail and dependent condition exists.

Timing of surgical debridement is a controversial topic in burn patient care¹³,¹⁴,¹⁵. Traditionally, a conservative approach with serial dressing changes, allows for the necrotic tissue to separate from the healthy wound bed which would later on be skin grafted. However, if this approach was prolonged, the constant release of pro-inflammatory factors would result in a systemic inflammatory state, which would aggravate the metabolic, immunologic, and systemic imbalance, leading to multiorgan failure and death. Furthermore, a greater delay in surgical debridement would often lead to the infection of the burnt areas, compounding the risk of death, and nefarious cicatricial problems such as hypertrophic scars or articular contractures that can impair a patient’s quality of life.

The presence of these scars, in aesthetically functional delicate areas may require further procedures to correct. Janzekovic¹⁶ described tangential debridement in 1970 and altered the burn patient surgical care paradigm. He proposed an earlier intervention with aggressive mechanical debridement and skin grafting¹¹ to reduce the wound exposure time, reducing the metabolic stress, infection rate, and therefore, complications and mortality rate. Additionally, length of stay and consequentially, costs are reduced¹³. This was particularly true in patients without airway lesion¹⁴.

Some authors still defend a more delayed approach, claiming that a higher blood loss and consequentially a higher need for transfusion leads to further metabolic and hemodynamic distress. Additionally, some authors defend that early burn depth is difficult to ascertain, making the distinction of which areas will spontaneously heal and which will require debridement and grafting a hard decision in the first days, even for experienced burn surgeons. This difficulty stems from the heterogeneity of a burn, where it is common that a patient presents with lesions with different prognoses in continuity and often in a spotted pattern; and the fact that burns are an evolving lesion. The zone of stasis is an area that presents a potentially reversible area, and adequate fluid therapy and infection prevention can greatly improve the outcome of this area¹⁵. In sum, those who defend a delayed approach suggest that deferring the surgical approach for some days will allow a more accurate assessment and prevent unnecessary interventions.

In our cohort, which replicates the current literature, we have observed a direct proportional relation between the surgical timing of the first intervention and the length of stay. (Figure 1). Patients who were submitted to an earlier intervention had a shorter length of stay. This opens an avenue for better burn care – an earlier approach may provide a shorter length of stay, which might lead to a decreased infection risk, particularly nosocomial infection, and a reduction of costs. This approach has been shown to overcome the benefits of a delayed surgical intervention¹⁷,¹⁸.

Figure 1 - Relationship between the first surgical intervention day and total length of stay.

Figure 1 - Relationship between the first surgical intervention day and total length of stay.

In our cohort, multiple factors have been associated with a delayed first surgical intervention such as patients who are critically unstable to tolerate a surgical procedure, patients who have multiple small areas that heal favorably with dressing changes, lack of operating room time, and the need to delay surgery due to the use of oral anticoagulants.

Burn patients are susceptible to infection, especially by nosocomial multidrug-resistant organisms. In our cohort, 46% of patients were diagnosed with at least one infection. The average length of stay in patients who had an infection was 40 days, which contrasts with 20 days in patients who never had a microorganism identified in admission cultures or required any further septic workup. This difference is even greater when nosocomial infections (defined as an infection that develops in the first 48 hours after admission), where the average length of stay was 46 days.

Interestingly, this increased length of stay was independent of the affected system (mucocutaneous, urinary, hematologic, or bronchial). There, infection prevention is one area where significant improvements can translate into reduced length of stay, furthering the cause for the development of specific strategies for infection control and prevention.

Some laboratory results reflect a worsening clinical status and are also associated with an increased length of stay. Low hemoglobin (defined as a laboratory value of hemoglobin under 8.0g/dL); Renal Insufficiency (defined by serologic creatinine over 1,2mg/dL); hypoproteinaemia (defined by serologic protein inferior to 60g/L) and hypoalbuminemia (defined by serologic albumin inferior to 35g/L) were all statistically significant to predict an increased length of stay.

Classic prognosis indexes, such as the Abbreviated Burn Severity Index and the Modified Baux Score, include known morbidity influencing factors such as surface burnt area and airway lesion. As stated previously, these factors also have a strong correlation with the length of stay.

Intra-hospital mortality in Portugal remains comparatively high (7,7%)⁷ when compared with other countries from southern Europe, but it is steadily decreasing. In our cohort, mortality was 6%. Burn mortality rates are normally calculated based on the general population that suffered a burn injury. This group is heterogeneous and includes small burn areas, lesser severe burn degrees, and reflects mostly patients who are mostly treated with dressing change in an ambulatory clinic or with a short infirmary stay. The mortality rate in our cohort reflects only patients who were admitted to a Burnt Patient Special Care Unit.

As a final note, the author would like to acknowledge some study limitations. First, the study is retrospective in design, which does not allow patient randomization. Secondly, some possibly important factors were not assessed such as the need for transfusion, microorganism resistance pattern, and what antibiotic therapy was realized. Functional and aesthetic outcome scales were not accessed. Bigger, multicentric studies might allow better stratification of patients according to their burn surface area or patient co-morbidities, which might be able to combat the heterogeneity of this specific patient population and allow a more practical conclusion that is applicable daily.

CONCLUSIONS

This study allows us to state that variables related to higher burn severity, such as burnt area, need for mechanical ventilation, need for fasciotomy or escharotomy, airway injury, and the presence of third-degree burns have a significant effect on the length of stay. However, in the author’s opinion, the most relevant conclusion in this retrospective study is the confirmation that modifiable factors exist – such as time to first intervention and the number of documented infections – that can effectively reduce the length of stay. These two areas should be the focus of patient care to improve health-related outcomes.

The conclusion of our study is on par with current medical literature.