Long-term Physical and Mental Health Outcomes of Older Adults Following a Major Burn Injury: A Burn Model System Investigation

Abstract

Older adults are at a higher risk of complications after burn injuries since many physical and mental changes are compounded by increasing age. Few studies have targeted the long-term effects of burns on older adults. Therefore, this study will investigate the long-term physical and mental health outcomes in older adults. About 3129 participants from the Burn Model System Database were divided into 3 cohorts based on their age at injury (18-54, 55-64, and 65+). Physical Component Summary (PCS) and Mental Component Summary (MCS) scores were derived from the 12-item Short Form (SF-12) and the Veterans RAND 12-item (VR-12) health surveys and analyzed to measure recovery at preinjury, discharge, 2-year follow-up, and 5-year follow-up. ANOVA, T-score analysis, and linear mixed-effects models were utilized to assess for significant differences in outcome scores. PCS scores were significantly different between the 18-54 cohort and 65+ cohorts at the preinjury and 2-year time intervals (P < .001 and P < .001, respectively) but not at the 5-year follow-up (P = .28). MCS scores were significantly different between the 18-54 cohort and 65+ cohorts at all time intervals measured (P = .001, P < .001, P < .001, and P = .005, respectively), though the change in MCS scores over time was not significantly different between age cohorts across time (P = .088). This supports that patients 65 years and older have a different physical function recovery trajectory when compared to patients under 64 years. These findings underscore the belief that for physical recovery after a burn injury, individualized physical rehabilitation plans will provide the most benefit for patients across all ages.

INTRODUCTION

Every year, it is estimated that approximately 600,000 people in the United States sustain burn injuries that require medical attention. Of those, approximately 50 000 require hospitalization.¹ The acute management of these injuries ranges from supportive care to surgical intervention, depending on the size and depth of the burn injury. Yet, even with the best acute care, burn injuries can have debilitating effects on a patient’s quality-of-life that can last for years after their injury.²

Physiologic changes after a major burn injury include fluid loss, a hypermetabolic state, and immunosuppression.³ These changes are managed acutely in the hospital setting. Less severe burns can be managed in an outpatient setting with frequent follow-ups to track recovery. In either case, chronic manifestations of the burn can present themselves weeks to months after the initial injury.⁴ These manifestations can negatively affect a patient’s physical and mental capabilities and reduce their quality-of-life. Physical effects can include decreased mobility, pain, limited function, loss of independence, and changes in appearance.⁵ Impacts on mental health can include new or worsening symptoms of depression and anxiety and an increase in sleep-related issues.⁶ Although the current literature is limited in investigating physical and mental changes beyond 2 years after the initial burn injury, a few studies suggest that physical and mental abilities can be impacted for up to 20 years or more.^2,6 The impact of these physical and mental changes intensifies with increasing age and the length of the acute burn treatment.⁶ As a result, not only do older patients have an increased risk of morbidity and mortality during the acute injury phase, but they also have an increased risk of complications in the postacute and chronic injury phases as well.⁴

As the United States population continues to age, increasing numbers of older adults (age > 55) with burn injuries will present to hospitals around the country.⁷ Very few studies have specifically examined the long-term effects of burns on older adults, limiting the knowledge that can guide management in this population.⁸ Understanding the physical and mental health impacts of burn injuries on older adults in the years after the initial injury is needed to manage long-term physical and mental health rehabilitation and recovery. Therefore, the purpose of this study is to investigate the trends in long-term physical and mental health outcomes up to 5 years postinjury in patients who were injured at 55 years of age and older.

METHODS

Study population

This study included adult participants who were enrolled in the Burn Model System (BMS) National Longitudinal Database from 1994 to 2022. The BMS is a program that began in 1994 and is funded by the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR), a Center within the Administration for Community Living (ACL), and the U.S. Department of Health and Human Services (HHS).⁹ There are multiple regional burn centers around the United States that participate in the BMS program, with each center providing comprehensive, multidisciplinary burn care for its region, including emergency medical care, postacute care, and long-term follow-up for patients with burn injuries.⁹ Data from each center’s patients were entered into a secure web application, REDCap, managed by the BMS National Data and Statistical Center based at the University of Washington,^10,11 where the information was deidentified and cleaned for analysis. Because of the deidentification process, no committee approval was required for this project. Patients aged 18 years and older who consented to participate in the BMS and provided data on any of the primary or secondary outcome variables of interest were included in this study. Participants were divided into 3 age categories based on their age at the time of their burn injury: age 18-54 years, age 55-64 years, and age 65+ years. The 18-54 cohort was used as the baseline comparator group to account for changes in the primary and secondary outcomes in younger patients. The older adult cohort was divided into a 55-64 cohort and a 65+ cohort in order to more precisely evaluate the relationship between increasing age and long-term functional outcomes, as described by Klein et al.,⁶ while providing a large enough sample size within age cohorts for appropriately powered statistical analysis.

Outcomes

The primary outcomes examined were the scores of 2 validated instruments: The 12-item Short Form Essay (SF-12) and the Veterans RAND 12-item score (VR-12). The SF-12 is a short questionnaire aimed at producing summary scores of a patient’s physical and mental health from the patient’s perspective. It is often used as one tool to measure a respondent’s self-reported quality-of-life.¹² In 2015, the BMS switched from using the SF-12 to the VR-12. Items on the VR-12 are like those on the SF-12, and the developers of the VR-12 created a crosswalk to convert VR-12 scores to the SF-12 metric. Using this crosswalk, VR-12 scores were converted to the SF-12 metric, and the combined SF-12 scores were used for all analyses. Scoring the SF-12/VR-12 results in a physical component summary (PCS) score based on physical function and a mental component summary (MCS) score based on mental function.

Secondary outcomes examined were scores from 3 other validated instruments: the Patient-Reported Outcome Measurement Information System (PROMIS) physical function score, the PROMIS pain interference score, and the Community Integration Questionnaire (CIQ). The PROMIS physical function score and the PROMIS pain interference scores are derived from 2 different 4-item short forms that measure physical function and pain, respectively.^13–15 The PROMIS physical function and pain interference scores have been normalized on a t-score metric and centered on the U.S. population with a mean score of 50 and a standard deviation of 10.^16–18 Finally, the CIQ focuses on the social and community integration of patients. The BMS only uses the CIQ-Social Integration (CIQ-S) subscale, which is a 7-item scale with scores ranging from 0 to 12. Higher scores denote better health for PCS, MCS, PROMIS physical function, and CIQ-S scores. In contrast, higher PROMIS pain interference scores denote more pain interference.

Primary and secondary outcomes were analyzed at 4 timepoints: preinjury (recalled, defined as a participant’s function in the month before injury, and collected at the time of study enrollment), discharge (from the acute, index hospitalization), 2 years after injury, and 5 years after injury. Each validated instrument had data collected only at certain timepoints by the BMS: SF-12 and VR-12 scores were collected at all timepoints, PROMIS physical function and pain interference scores were only collected at the 2- and 5-year follow-up timepoints, and the CIQ-S scores were collected at preinjury, 2 years, and 5 years. Individuals with missing data from the primary and secondary outcomes were excluded from the analyses.

Statistical analysis

An analysis of variance (ANOVA) was utilized to compare scores on the outcome measures across the 18-54, 55-64, and 65+ age cohorts at different timepoints. If the ANOVA analysis indicated that a statistically significant difference between the cohorts existed, then a post hoc analysis was completed to determine which 2 age cohorts had statistically significant differences between their scores using the Student’s t-test. In addition to the ANOVA analyses, a linear mixed-effects model was used to analyze PCS and MCS scores to determine if there were differences in scores over time between age cohorts. The advantage of mixed-effects models is that they allow for the simultaneous modeling of repeated measures while accounting for correlations of measures. This allows us to assess longitudinal trends, including differences in trends between cohorts. The models included a fixed effects indicator for the age cohort, an indicator for follow-up timepoints, and an interaction between the age cohort and timepoint indicators, as well as sex and the corresponding preinjury score. A random intercept was included for each participant. An omnibus test was then used to determine if the interaction terms were significant. A significant finding indicates that the change in scores over time differs across the age cohorts. For all statistical analyses, a P < .05 was utilized to indicate a significant finding. All analyses were completed using Stata v.18 (StataCorp. 2023. Stata Statistical Software: Release 18. College Station, TX: StataCorp LLC.).

RESULTS

A total of 3129 individuals participated in this study (Table 1). There were 2389 participants in the 18-54 cohort, 431 participants in the 55-64 cohort, and 308 participants in the 65+ cohort. 71% of the studied population were male, and the average burn size (% total body surface area (%TBSA)) for all participants was 17.9. Racially, most of these patients were white, followed by black/African Americans. The most common type of burn injury across all age groups was fire/flame burns, and the median length of stay was 28.2 days. Postrecovery, most of the patients in the dataset were discharged home. However, in the 65+ age cohort, there was a larger percentage of patients who were discharged to an extended care facility when compared to the other 2 age cohorts.

PCS scores (Tables 2–4)

There were significant differences between all 3 age cohorts at preinjury recall (F(2, 2639) = 59.9, P < .0001), with the 55-64 and 65+ age cohorts reporting lower physical health (t = −7.95, P < .001 and t = −8.50, P < .001, respectively). A similar result was found at 2 years postinjury (F(2, 1445) = 12.2, P < .0001), with the 55-64 and 65+ age cohorts reporting lower physical health than the 18-54 cohort (t = −4.15, P < .001 and t = −3.33, P = .003, respectively). However, no significant group differences in PCS scores were found at discharge and 5 years (F(2, 1877) = 1.35, P = .25 and F(2, 292) = 1.29, P = .28, respectively). The results of the mixed modeling also indicate that the different age cohorts may recover at different rates, as the omnibus test for the interaction term (age cohort × time) was statistically significant (X²(4) = 19.65, P = .0006).

MCS scores (Tables 5–7)

There was a significant difference between the 3 age cohorts at all timepoints measured (F(2, 2639 = 6.52, P = .002); F(2, 1877) = 12.02, P < .0001; F(2, 1445) = 14.13, P < .0001; F(2, 292) = 5.2, P = .006 for preinjury, discharge, 2 year, and 5 year, respectively). The 18-54 cohort had significantly lower mental health scores at all timepoints when compared with the 65+ cohort (t = 3.59, P = .001; t = 4.73, P < .001; t = 5.27, P < .001; t = 3.15, P = .005 for preinjury, discharge, 2 year, and 5 year, respectively). Similarly, the 55-64 cohort also had lower MCS scores than the 65+ cohort at discharge and 2 years (t = 2.55, P = .029; t = 4.24, P < .001, respectively). However, when comparing the trajectories in MCS scores using mixed modeling, the omnibus test for the interaction term (age cohort x time) did not support that there were differences in the rate of recovery between age cohorts across time (X²(4) = 8.11, P = .088).

Community Integration Questionnaire-Social Integration (CIQ-S) (Table 8)

There was a significant difference in scores between the 3 age cohorts at the preinjury timepoint (F(2, 7534) = 40.63, P < .0001). At this timepoint, there was a significant difference between the 18-54 cohort and the 55-64 cohort (t = −7.89, P < .001), as well as between the 18-54 cohort and the 65+ cohort (t =−5.50, P < .001), with the 18-54 cohort having higher CIQ-S scores than the 2 older cohorts. However, there were no significant differences in CIQ-S scores between the 3 age cohorts in the 2 other timepoints measured (F(2, 1456) = 1.16, P = .31 for 2-year follow-up and F(2, 146) = 0.03, P = .96 5-year follow-up).

PROMIS physical function (Table 9)

There was a significant difference in scores between the 3 age cohorts at the 2-year follow-up timepoint (F(2, 488) = 10.65, P < .0001) but not at the 5-year follow-up (F(2, 154) = 1.1, P = .34). Further analysis at 2 years showed that there was a significant difference between the 18-54 cohort and the 55-64 cohort (t =−3.50, P = .001), as well as between the 18-54 cohort and the 65+ cohort (t =−3.78, P = .001), with the 18-54 cohort having higher scores than the 2 older cohorts.

PROMIS pain interference (Table 10)

ANOVA analysis of PROMIS Pain interference scores found no significant differences in scores between age cohorts at either the 2- or 5-year follow-up (F(2, 484) = 2.42, P = .089 and F(2, 153) = 1.56, P = .21, respectively).

DISCUSSION

This study investigated trends in long-term physical and mental health outcomes at discharge, 2 years, and 5 years postinjury in participants 55 years of age and older who sustained burn injuries. The data shows that although there was a significant difference in PCS scores between the youngest and oldest cohorts at the 2-year follow-up, there were no significant differences in PCS scores between any age cohort at the 5-year follow-up. Second, there was a significant difference in the recovery trajectories in PCS scores between age cohorts. Third, MCS scores were significantly lower for the 18-54 cohort compared to the 65+ cohort at all timepoints, though the rate of mental health recovery was not significantly different between the 3 age cohorts.

Physical component summary

PCS is a composite score that incorporates measures of physical function, limitations caused by physical problems, bodily pain, and general health. A higher PCS score represents better physical health.¹⁹ A previous study by Stephens et al. estimated that average PCS scores are approximately 52 for age 55-59, 51 for 60-64, and 49 for 65+ in the general population.²⁰ To further characterize the data, minimum clinically important difference (MCID) can also be used to provide context for the minimum PCS change that would be seen for a clinical difference in a patient’s physical function. MCID for PCS in people who have experienced burn injuries has not been established, though there have been other studies looking at the MCIDs for PCS for other forms of trauma. One study looking at MICD for stroke patients found that the MCID for PCS was between 1.8 and 3.0.¹⁶ Another study looking at patients with cervical spondolytic myelopathy describes an MCID greater than 9.6 for PCS as clinically significant.²¹ The change in mean PCS scores within this study was greater than 10 points for each age cohort between discharge and the 5-year follow-up, which is higher than the previous studies described. This indicates that the physical recovery within each age cohort is both statistically and clinically significant. However, because the difference in PCS scores between the age cohorts was <5, it is difficult to adequately conclude if there are clinically significant differences in the physical quality-of-life between age cohorts, even if they were statistically significant.²²

Typically, we associate older patients with a lower PCS score when compared with younger patients.²³ This was the case in the earlier PCS scores and is the pattern that was observed in this study, with the 65+ cohort having a significantly lower score than the youngest cohort preinjury and at the 2-year follow-up. However, when looking at PCS scores at the 5-year follow-up, there is no significant difference in scores between the 3 age cohorts. There is also an upward trend in PCS scores in all age cohorts across time. In addition, the linear mixed-effects model showed that participants in the 65+ age cohort recover at a different rate when compared with the 18-54 cohort at the 2-year follow-up. Unfortunately, the model does not determine if the 65+ age cohort recovers at a faster or slower rate. However, the data showed that at the 5-year follow-up period, the 2 older cohorts did not have significantly different PCS scores when compared to the 18-54 cohort, hinting that the older adults’ physical quality and function caught up with their younger counterparts in the years following their burn injury. This further supports the idea that older patients recover differently than younger patients physically.

One explanation for this is that although older patients’ physical health deteriorates more with serious injury, it does not mean that they are going to recover slower than their younger counterparts. In fact, this could support that older adults could maintain their current recovery velocity for much longer than younger patients since previous studies have shown that older adults can benefit greatly from physical activity and physical therapy services, especially in the recovery and maintenance phases of their physical health.²⁴ Though, this could be confounded if older adults have different expectations of their own abilities compared to younger adults, leading to different self-rated scores, thus skewing the data. This is known as response shift.²⁵ Older adults usually have lower physical quality and function when compared to younger adults and thus are more accustomed to having a lower physical quality and function, even in the face of a major injury. However, younger adults may not be accustomed to having a lower physical function and may imagine the impact of an injury to be larger than what an older adult may perceive. As a result, the true PCS scores between young and older adults may be different than what is reported on the dataset and could account for the results of this study. Another explanation is that there are more resources currently dedicated to helping older adults recover than younger patients. Most older adults remain insured under Medicare, which incentivizes services that promote and improve the delivery of care.²⁶ As older adults require more time in physical rehabilitation and other related services in the first place, this exposes them to higher quality of care for a longer period, which builds on each other to help older patients physically recover until their perceived physical health is similar to their younger counterparts.

Mental component summary

Another interesting trend was identified in the review of the MCS scores. The MCS score is calculated from questions based on mental health, psychosocial disability, and social adjustment. One study estimated that the average MCS scores are approximately 50.5 for ages 55-59, 51 for 60-64, and 52 for 65+.²⁰ As with PCS, MCID for MCS has not been established for burn patients. However, one study looking at critically ill patients estimated that the MCID for MCS is 3.²⁷ The data within our study does not show that MCS scores reached this threshold within the age cohorts between discharge and the 5-year follow-up. However, between age cohorts, there was a minimum 4-point difference between age cohorts, indicating a clinical and statistical significance in MCS scores.

When looking at MCS scores, the data did not show that older participants recovered the quality of their mental health slower than their younger counterparts. Although there were expected age-related differences in MCS scores between the younger and older age cohorts, the data shows that older participants’ ability to recover mentally was not associated with their age cohort or with the follow-up interval. In fact, the data shows that the 65+ cohort had higher average MCS scores than the other 2 age cohorts across all time intervals. This supports the fact that although older patients may be more fragile physically, their mental resilience is not significantly different from that of their younger counterparts.²⁸ Another explanation could be that even though older adults may receive more resources than their younger counterparts, older adults might recover better with similar amounts of resources when compared to younger patients. In fact, a previous study showed that older adults can recover better than younger adults utilizing psychological therapy, especially with anxiety and depression disorders.²⁹

Secondary outcomes

When analyzing the secondary outcomes, the PROMIS physical function scores showed similar results when compared to PCS scores as discussed earlier: at the 2-year follow-up, younger patients reported higher physical quality and function than older patients, but not at the 5-year follow-up. However, when analyzing PROMIS pain scores, there were no significant differences between age cohorts at the 2-year and 5-year follow-up. This can signify that in the long-term follow-up period, patients’ perceptions of pain remain similar across the studied age cohorts. CIQ-S scores were not significantly different between age cohorts at the 2 follow-up periods. This can signify that regardless of age, burn patients feel the same way about how well they are able to reintegrate into their community in the long term. In addition, the CIQ-S scores were on the higher end of the scoring spectrum, indicating that burn patients felt that they were able to reintegrate into their community well.

Limitations

There were several limitations to this study. The most significant was the small sample size, especially in the older age cohorts. As the follow-up time interval increases, the attrition rate also increases and reduces available participants for analysis. For example, there were 256 participants in the 65+ cohort with PCS/MCS scores at the preinjury time interval. However, that number dropped to just 46 participants by the 5-year follow-up interval. This large discrepancy in sample size over time can introduce attrition bias into the study, causing the older age cohorts to be underrepresented in the study, especially at later follow-up intervals. Consequently, this can limit how this study can be generalized to how older adults recover in the long term, given that the data is slowly skewed towards younger patients as time progresses. In fact, this study initially attempted to have a 75+ age cohort in the analysis, but a small sample size within that age cohort prohibited properly powered statistical analysis for that age cohort.

Another limitation to recognize is that this study was unable to include other factors that can influence functional status, such as %TBSA and hospital length of stay, without compromising statistical power.³⁰ Early designs of this study attempted to stratify the participants by these variables, but analysis showed an inadequate sample size for properly powered analysis, especially for the older age cohorts and at the 2- and 5-year follow-up periods. As a result, analyses stratifying by these variables were deferred to a future study once more data could be collected. Additionally, the limited sample size prohibited mixed model analysis on some of the study variables and ad hoc analysis of significant differences from the ANOVA analyses, especially at later follow-up periods. Other measures, such as the preinjury and discharge scores for the PROMIS physical function, PROMIS pain, and CIQ-S scores, were not collected, which prevented analysis at these timepoints.

In all, the main limiting factor of this study was the small sample size at the older ages and longer follow-ups, which is attributed to the high attrition rate. This high attrition rate can be attributed to a variety of factors. One factor is that participants are randomly lost to follow-up. In the BMS protocol, follow-up is conducted by contacting participants at predetermined intervals, either by phone or by mail.³¹ If patients move or simply change their contact information, it can be difficult to find these participants for long-term follow-up, especially years out from injury. This attrition bias is likely independent of the participant’s recovery status, but it is difficult to determine if long-term functional status plays a role in this type of attrition. However, it is difficult to ignore that the attrition was unequal between the cohorts and follow-up periods. One reason for this could be that those who continue to follow-up in the long term are those patients who have continuing health concerns from their burn injury, whereas those patients who are lost to follow-up can be “satisfactorily recovered” or “too healthy” to follow up in the long term. The effect of this unequal attrition was that as time progressed, the data was slowly skewed towards those participants with lasting debilitation from their injury. Although this provides us with a more focused pool of participants for analyses of those who had a slower recovery, it limits the generalizability of this study to those patients who had a better recovery trajectory and for the general population. Yet, it is important to recognize that this study only analyzed data from one nationwide database. Future studies could incorporate multiple databases to increase the sample size for a more thorough analysis.

CONCLUSION

Our data demonstrates that older patients have a lower self-reported physical health over time when compared to patients under 54 years up to 2 years postinjury. However, at 5 years following injury, there was no significant difference in self-reported physical health between age cohorts. Mental health scores were higher in participants aged 65+ years, though the recovery rate was not significantly different between age groups across time. These findings underscore the belief that for physical recovery after a burn injury, a one-size-fits-all plan will not provide maximum benefit to burn patients of any age, and the physical rehabilitation plan should be individualized.

Studies need to be conducted in the future to further characterize how older adult patients recover in the years following their injury. For example, although we have enough data to characterize how older adult patients recover 1-2 years postinjury, it would be beneficial to conduct a study on how older adult patients recover beyond 5 years after their injury. Once more data has been collected, this analysis could be repeated with a larger sample size. Other measures of physical and mental health could be included in future studies as well to further characterize differences in mental, physical, and social reintegration. Until then, we must be vigilant in closely monitoring how our older population reacts after a stressful and traumatic event and respond appropriately to ensure the best quality of care for our increasingly aging population.

Funding:

The contents of this abstract were developed under a grant from the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR grant number 90PBU0006). NIDILRR is a Center within the Administration for Community Living (ACL), Department of Health and Human Services (HHS). The contents of this abstract do not necessarily represent the policy of NIDILRR, ACL, or HHS, and you should not assume endorsement by the Federal Government.

Conflict of Interest Statement:

None declared.

REFERENCES