The outcome variables that have been examined in previous studies include primary health-care outcomes (death, complications, disposition, functional status) and surrogate or secondary health-care outcomes (length of stay, ICU length of stay, cost of care, readmission).

Many of the most widely quoted studies examining outcomes in GTP have been retrospective and have utilized institutional, state, and national data sets in order to accumulate large numbers of patients [2–7]. Most have focused upon patient-specific criteria present on admission to the hospital that could be measured, abstracted, and included in registry data sets. These studies have provided useful descriptions of the association between PECs, age, ISS, physiology at presentation, and outcomes such as mortality, inhospital complications, functional status at discharge, length of stay, and discharge destination. The overriding theme of these studies is that most primary health-care outcomes are negatively affected by PECs and that more PECs per patient generally portend a worse outcome. Specific PECs such as preexisting cirrhosis and kidney failure [2, 5] are associated with higher odds of death for any given severity of injury. Several studies have also shown that the impact of PECs may be most important in GTPs with moderate injury (ISS 10–20) as opposed to mild or severe injury [2, 8].

Many of these studies also demonstrated that age alone is an independent predictor of outcome and that aggregate injury severity and deranged physiology on admission are also associated with worse outcomes. Severe traumatic brain injury or low GCS on admission in the elderly [3, 6] has been consistently associated with poor outcomes.

Intuitively, none of these results seem surprising, but favorable outcomes observed for some seriously injured GTP [7, 9, 10] and the feasibility of performing major surgery in geriatric cohorts [11] have led investigators to pursue more precise predictors of outcome. The underlying concept that physiologic reserve is not necessarily predicted by age or even PECs has been accepted. Following trends in general geriatrics research, more recent studies on patient-specific factors affecting outcomes for EGS and GTP have focused upon determination of frailty [12–14] or functional independence [15, 16] in predicting the primary and secondary outcomes described above.

Frailty and functional dependence are related but separate concepts that have been used as theoretical and practical constructs to describe patient-specific characteristics in the elderly and “non-elderly” population. While a complete discussion of the topic is beyond the scope of this chapter, several reviews have been published [13, 17]. What is consistently clear is that frailty and functional dependence have a measurable negative impact on all outcomes including mortality, complications, length of stay, and discharge disposition [11–13].

Some assessments of frailty include subjective criteria, while others are more “rule based” including timed maneuvers such as the “get up and go” assessment or tests of grip strength [14, 17]. While rule-based assessments may be difficult in the acute hospital setting, several recent studies have confirmed that it is possible [12, 18] to perform them. It has also been shown that these scores may be modified to provide rapid assessment for trauma patients for the purpose of prognosis, especially with regard to discharge disposition [13, 18]. Inhospital tools for assessment of functional dependence that relies principally on historic factors [13, 14] are based on the Karnofsky scale or simplified equivalents. These tools are easily reproduced and can also be used for prognostic purposes especially with regard to disposition.

Given the recent emphasis on the importance of assessments of frailty and functional dependence, there are some issues that exist that may limit their routine use in clinical practice for acute care surgeons. Many assessment tools require interaction between assessor and patient, and, where patients are unwilling or unable (dementia, obtundation, language barrier) to participate, a surrogate must be available who can provide accurate, unbiased information. One recent study [13] reported a large number of exclusions related to the surrogate, and of the patients enrolled, more than half relied on the surrogate alone for patient information. In our own experience, family members may overestimate, underestimate, or disagree with respect to some of the variables assessed. Another potential limitation of frailty assessment is that in the outpatient setting there may be opportunities for modification via exercise programs to enhance muscle mass, or management of depression [14], it is less clear that interventions can be undertaken in the acute, inpatient setting. Finally, unlike the elective surgical setting where it might be possible to use these data in making decisions regarding patient selection for surgery, it is less possible to do so under emergent circumstances.

Given that patients will present themselves to the health-care system with a set of patient-specific factors, there are some attributes of the system that may help in determining primary and secondary health-care outcomes. The system may be broadly defined as a continuum from the macro (state and regional delivery systems) to the micro (specific features of care delivery within institutions).

This continuum of care has been evaluated with respect to outcome for GTP. These data are generally lacking for EGS although studies have been done that suggest emergency general surgery outcomes for certain index cases are improved using the “Acute Care Surgery model” [19]. These studies suggest that a verification process might be warranted for hospitals performing emergency general surgery and by inference it is possible that EGS patients would fare better in such institutions. To our knowledge this hypothesis has not been directly tested for elderly emergency general surgery patients.

Several studies have demonstrated that GTP may be under triaged [20, 21] based upon existing field triage criteria (CDC). The American College of Surgeons Committee on Trauma has recommended alteration in field triage guidelines to improve recognition of injury in patients over the age of 55 [22]. These include using a systolic blood pressure of 110–120 [23] as opposed to 90 to define hypotension in the elderly and therefore establish a higher (presumably more sensitive) threshold for trauma system activation based upon “Box 1” physiologic criteria. For centers utilizing a multitiered response, the highest tier of trauma system activation would be used more often in GTP.

A significant proportion of GTPs are cared for outside of trauma centers or in centers with lower levels of designation [24–26]. Some studies have suggested that GTPs with both major and minor injuries have better outcomes in trauma centers than non-trauma centers [26, 27], while another recent study [28] demonstrated equivalent outcomes for all trauma patients with minor injuries treated in non-trauma centers as compared to trauma centers. This is a complex issue for several reasons. First, as suggested by regional data in New York state [24], the majority of trauma patients treated in non-trauma centers is elderly and has low injury severity, but these hospitals do not participate in New York state or national trauma registries. Other than mortality, precise outcome data for patients treated in these hospitals is lacking, and it may be inaccurate to assume they are better or worse than in trauma centers. Second, inclusion or exclusion of isolated hip fractures [29] in the registries of trauma centers is highly variable and may bias outcome data for individual centers so that overall survival data between institutions is not comparable. These patients have lower ISS and lower overall rates of complications than age-matched controls with multisystem trauma. Third, and along these same lines, the definition of “trauma” has expanded to include isolated single-system orthopedic injury (hip, upper extremity), many of whom are admitted to trauma centers and included in their registries. It is important to recognize that even within trauma centers, a significant proportion of these patients may be admitted to non-trauma services. This has not been well quantified in studies that suggest admission to trauma centers is beneficial for GTPs. As noted above, such patients tend to have more comorbidity and lower ISS. It may not be appropriate to view such patients as under triaged based simply upon the presence of an injury in a large administrative data set when in fact the principle reason for hospital admission was not the injury but the medical conditions associated with it. These distinctions can be difficult to make in reviewing large databases. Finally, on an intuitive level, the impact of removing a GTP with minor injury from their primary care network to a major trauma center where their medical history is not well known is difficult to measure. Is care less efficient? Do such patients receive unnecessary testing due to lack of familiarity with their preexisting conditions? Are patient-centered values as well understood by treatment teams who may have had no prior relationship with the patient or family? What does seem reasonable is that the triage threshold for elderly patients with any physiologic or anatomic criteria should be lowered and that such patients should be treated in trauma centers. This reasoning should probably be extended to patients in “Box 3” (mechanism of injury) but may not always apply for ALL elderly patients who sustain same-level falls with isolated extremity fractures.

Recently, the concept of “mandatory” trauma center care for GTPs has begun to focus more upon the value of a concentrated experience in geriatric trauma as defined by the proportion of GTPs seen relative to total volume [30, 31]. These studies used state and national data to suggest that the benefit of trauma center care is most pronounced in centers that care for a higher proportion of geriatric patients as opposed to those with high overall volume. One study [31] reports a reduced risk-adjusted mortality in these high-proportion centers which more often tended to be lower-volume, community Level II centers. This study calls for the regionalization of geriatric trauma and creation of geriatric trauma centers.

Within hospitals or trauma centers, recent studies have focused upon geriatric consultation and geriatric trauma services as possible means of improving outcome and process metrics. Complication rates including delirium and urinary tract infection are improved when GTPs are admitted to geriatric trauma services or receive geriatric consultation [32–34]. Process of care including discharge disposition is also enhanced. Differences in mortality rates have not been demonstrated using these measures. In her recent review of geriatric trauma, Kozar [35] proposes that “ …any significantly injured geriatric trauma patient should be admitted by the trauma surgeon with appropriate consultation and multidisciplinary input.” This seems to be a reasonable suggestion and allows interpretation and “local rules” to determine where GTP with isolated single-system orthopedic injury will be managed. When available, geriatricians would seem to be an excellent resource, but their presence and availability is limited and likely to remain so. Identifying a hospitalist or internist who can work closely with the trauma team on a consistent basis might be another option.

Discharge disposition is a systemic factor that deserves special consideration because it has been used by many authors as a surrogate [2–4, 33, 34] outcome and quality metric. Most studies have defined discharge to home or acute rehabilitation centers as “favorable,” while death or discharge to a skilled nursing facility is “unfavorable.” Disposition is a primary health-care outcome that patients know and care about since it is directly related to the level of functional independence. It can be argued that this is one of the more important outcome variables for GTP, but relatively little is known about it. In Kozar’s [35] review, a questionnaire distributed largely to academic trauma surgeons indicated that there was strong agreement with the statement that “post discharge rehabilitation is a major issue in the long term outcome of GTP.” There are confounding variables that affect post-discharge planning. A recent study examining national data [36] identified clear trends in the reduction of acute rehabilitation beds available for trauma patients and links this trend to the evolution of stroke center certification that closely tracks disposition to acute rehab as a requirement for certification. A recent review of effectiveness of rehabilitative services for traumatic brain injury [37] showed that elderly patients with TBI have more limited access to acute rehabilitative services based upon age alone and typically end up in acute rehab centers only when their cognitive scores are higher than younger TBI patients. These considerations regarding disposition as an outcome suggest that there may be many other factors determining disposition for GTP that are independent of patient factors or the quality of acute inpatient care. It should be an area of focus for geriatric trauma research.

The outcome of treatment for TBI in the elderly is an especially important example of a specific condition requiring careful outcome analysis since it is a common form of injury in GTP. Canadian Health Ministry data from 2006 to 2011 [40] demonstrate a 24 % increase in admissions to Canadian hospitals for TBI in patients over 65 years of age. Furthermore, these hospitalized TBI patients demonstrated an increased Charlson Comorbidity Index (CCI) over the period of review. While patients with severe TBI and low GCS are known to have very high mortality and likely have a high prevalence of care withdrawal, mild and moderate TBI are less well understood [3]. A recent meta-analysis [41] demonstrated an inhospital mortality of 12.3 and 23.3 %, respectively, for mild and moderate brain injury in GTP. For both groups, mortality increased at 6 months, and for both measures, it was higher than younger counterparts. Not surprisingly, there was a strong correlation between preexisting conditions/chronic health issues and mortality. The management of both acute and chronic subdural hematomas (SDH) has been studied and reported in the neurosurgical literature since the 1960s. The standard of care for younger patients with acute SDH has been evacuation within 4 h [42]. This standard has been challenged in recent years and may be inappropriate in the elderly given the presence or absence of mass affect, underlying brain injury, neurologic exam, and presence of anticoagulants.

Elderly trauma patients have a higher proportion of extra-axial mass lesions (epidural and subdural hematomas) than younger patients, and this trend begins at age 50 [43]. Data from the University of Michigan [44] demonstrate an inhospital mortality of 27/103 (33 %) for patients > 80 years of age undergoing evacuation of acute SDH. For patients between 50 and 79 years of age, mortality was 36/103 (22 %), not significantly different than patients under the age of 50.

Chronic SDH is a very different entity than acute SDH and may be a blended condition with both elements present. The classic description of a chronic SDH involves cognitive and physical decline with or without an evolving focal motor or speech deficit. There may be NO antecedent history of trauma. It is common for patients to present with recent falls and “acute and chronic SDH.” Miranda [45] reviewed 209 cases of chronic SDH seen over an 8-year period. One hundred thirty-nine (66 %) underwent some form of operative decompression, and 72 (34 %) were observed. Overall inhospital mortality was 35/209 (16.7 %); 44/209 (17.5 %) were discharged home. Median survival for the entire group over the period of follow-up was 4.4 years. Options for operative treatment of chronic SDH include burr holes or craniotomy; one recent review and meta-analysis suggests burr holes are favored when possible. This makes intuitive sense and factors into decision-making regarding treatment of acute SDH. When possible, neurosurgeons may wish to delay operative treatment of acute SDH in hopes that the hematoma will liquefy or become more “chronic” in nature so that a less invasive surgical approach may be used.

It seems clear that there is acceptable mortality associated with operative treatment of acute and chronic SDH, but studies have not clearly defined whether operative versus nonoperative therapy is preferred for these patients. The use of observational therapy for SDH has been described since the 1970s [46]. The simple presence of an SDH is not sufficient to warrant emergent evacuation and must be weighed against the likelihood that neurologic deficits will be improved or reversed by surgery. These considerations must be balanced against the risks of bleeding and hematoma recurrence associated with antiplatelet or anticoagulant medications. In summary, the management of acute and chronic extra-axial mass lesions in GTP requires careful judgment and patient selection.

We know that the inhospital mortality associated with mild and moderate brain injury of all types is under 25 %, but we do not have a good sense of the impact of such injury on functional outcomes. Good quality data on functional outcomes after TBI in the elderly is generally lacking, but a recent publication of comparative effectiveness research in TBI rehabilitation [37] sheds light on a few important concepts. Patients entering rehab centers with higher cognitive FIM scores tend to be older, but older patients also tended to have lower FIM scores 9 months post discharge. Older patients with TBI were less likely to return to independent living, had more returns to acute care hospitals from rehab, and had more readmissions to acute care hospitals during the 9-month follow-up period. The data reveal that a low proportion of elderly patients with severe TBI enter rehabilitation units. Only patients with higher cognitive FIM scores are likely to enter brain injury rehabilitation, but they tend to fare worse than younger counterparts with respect to standard rehabilitation outcome metrics.

Fractures of the cervical spine are common in elderly patients. In a descriptive study covering 5 years at a busy Level I trauma center, 139/726 (19 %) of cervical spine fractures were in patients over the age of 65 who were injured after ground-level falls [47]. Cervical spine fractures in GTP are most commonly of C1 or C2 (27 and 53 %, respectively), whereas these fractures account for only 13 and 21 % of cervical fractures in a younger cohort. Issues affecting outcome related to upper cervical fractures in the elderly include operative versus nonoperative therapy and the relatively high prevalence of dysphagia and aspiration complicating this condition.

Several retrospective series have addressed the prevalence of complications [48–50], associated with cervical immobilization in rigid collars and/or halo vests. The incidence of pneumonia was 10–14 % for these devices, and mortality ranged between 8 and 13 %. A more detailed retrospective review [51] was conducted examining operative versus nonoperative treatment of C2 fractures. It demonstrated 14 % 1-year and 44 % 2-year mortality for all patients, but slightly higher 30-day mortality for nonoperative treatment. Length of stay was significantly higher in the operative group (15 vs 7 days), and a greater number of operated patients required feeding tubes (18 %) as a result of dysphagia. Based on these retrospective data, the investigators conducted a prospective study of odontoid fracture management in an effort to define the role of operative versus nonoperative therapy and the complication profile associated with each [52–54]. In this non-randomized series, 156 patients were followed for 12 months. 101 underwent surgery by a group of dedicated spine surgeons; 58 had nonoperative treatment in cervical collars. Mortality for all patients was 18 % at 12 months. Dysphagia was more common (11 vs 5 %) in operated than nonoperated patients. The rate of all complications was 30 and 38 % in nonoperated versus operated patients. Disability related to neck pain was less severe in patients undergoing surgery versus those who did not undergo surgery. Selection bias was acknowledged as a potential issue in this prospective study.