Elder Mistreatment Documentation by Prehospital Clinicians: An Analysis of the National Emergency Medical Services Information System Database

Abstract

Objectives:

Elder mistreatment (EM), encompassing abuse and neglect, is a significant public health issue, affecting up to 10% of community-dwelling older adults annually. Elder mistreatment is a growing concern with a higher prevalence in institutional settings and substantial associated healthcare costs. Prehospital clinicians (PHCs) such as emergency medical technicians and paramedics are uniquely positioned to detect and report EM during their interactions with older adults in their homes. The objective of the study is to describe the rate and characteristics of EM documented by PHCs using the National Emergency Medical Services Information System (NEMSIS) database.

Methods:

This study analyzed data from NEMSIS, which includes standardized information about PHC emergency response encounters across the United States. In 2018, 22,532,890 activations were included from 9,599 agencies in 43 states and US territories. Elder mistreatment was identified using specific International Classification of Diseases (ICD) codes related to EM. Demographic data, injury location, and associated physical findings were also examined.

Results:

Out of 9,605,522 EMS encounters for patients aged ≥60, EM was coded in 1,765 encounters (0.02%). Most EM cases were listed as the cause of injury (64%), followed by the clinician’s first impression (25.4%). Physical abuse was the most common type of mistreatment reported (20.8%), followed by sexual abuse (18.2%), neglect (9.7%), and psychological/emotional abuse (0.34%). The median age of patients with documented EM was 72, and 62.3% were female. The most common anatomic locations of injuries were the lower extremities, head, and upper extremities.

Conclusions:

Despite the high prevalence of EM, PHCs infrequently document EM in their encounters with older adults. Additional training and comprehensive protocols are needed to improve the identification and reporting of EM, mainly elder neglect. Empowering PHCs through education and protocol development can significantly impact the detection and intervention of EM.

Introduction

Elder mistreatment (EM), which comprises elder abuse and neglect, is a significant public health issue, affecting as many as 1 in 10 community-dwelling older adults each year (1, 2). Those living in an institution are at even higher risk, with 64% of institutional staff admitting to perpetrating EM within the last year (3). In addition to impacting quality of life, this mistreatment is associated with adverse health outcomes, including depression, exacerbation of chronic illness, and increased hospitalizations (4). The excess direct medical costs associated with elder mistreatment have been estimated at $5.3 billion (5). This will likely become a more significant problem as the older adult population grows. Most EM goes undetected and unreported, which contributes to the impact of the phenomenon (6). Elder mistreatment is often ongoing, with multiple separate incidents occurring over time (7). Therefore, improving detection is essential to facilitate intervention to break the cycle of mistreatment and associated adverse outcomes.

Prehospital clinicians (PHC), such as emergency medical technicians (EMTs), paramedics, and first responders often associated with local/regional fire departments, who are front-line healthcare clinicians that typically treat patients in their own homes, are in a unique position to detect and report EM (6, 8, 9). This will not be the case in life-threatening situations, as the priority is stabilization and transport of the patient. However, upon arriving at the scene for non-life-threatening calls, the Emergency Medical Technician (EMT) follows established protocols to assess the patient’s condition, including vital signs and any chronic issues. In addition to providing medical care, the EMT offers basic social support and checks other aspects of wellness, adhering to guidelines that ensure consistent and comprehensive patient evaluation. The EMT ensures the patient’s stability, referring to additional services when necessary, and prepares them for transport to a healthcare facility for further assessment.

Older adults (aged ≥65) represent one in three Emergency Medical Services (EMS) activations (10), and many older adults receive multiple PHC assessments during a short period (11). Older adults experiencing EM are nearly three times as likely to utilize emergency services as other older adults (12). Prehospital clinicians can see inside patients’ homes unlike most other healthcare clinicians. They are trained to identify and report on environmental issues, such as lack of food, extreme temperatures, infestations, and excessive clutter, which may represent unsafe living conditions (9). They may also witness and report on specific interactions between a patient and caregivers or family members. Combined, these factors may suggest mistreatment, such as caregivers who are neglectful or unaware of the patient’s needs or patients who are afraid to talk in front of the caregiver. Further, if an older adult refuses transport to the hospital after 9–1-1 activation, PHCs are the only healthcare clinicians who assess them (13). Prehospital clinicians are mandatory reporters of EM in nearly all US states and have a professional responsibility to report their suspicions to Adult Protective Services (APS).

Despite their unique opportunity to identify and report this common and severe phenomenon, little is known about current PHC practice around EM. To date, many PHC documentation protocols do not include elder mistreatment as a focused data point to assess, and no studies have examined how commonly PHC document concerns about EM as part of providing care to older adults. The present study aimed to describe the prevalence rate and characteristics of EM, as documented by PHCs, and later encoded into the record, using the National Emergency Medical Services Information System (NEMSIS) database (available at NEMSIS.org).

Methods

Data Source, Study Design, and Study Population

Data were drawn from the National Emergency Medical Services Information System. The NEMSIS collects, stores, and shares standardized data about PHC emergency response encounters across the United States. The NEMSIS database contains variables on all aspects of the EMS encounter, including, but not limited to, information on the agency and unit responding, dispatch times, and details about the patient, scene, and situation within which a service was provided. It also includes vital signs, procedures performed, and the final disposition of the visit. This information is from the PHC’s documentation of each clinical encounter (which is typically electronic). We included 22,532,890 total activations from 9,599 agencies in 43 states and US territories from NEMSIS collected in 2018 to avoid potential bias from the COVID-19 pandemic. This included 87% of all EMS activations during 2018 in the US (14). This study was approved by the institutional review board at Weill Cornell Medicine with a waiver of informed consent.

Measures of Interest

Aspects of PHC documentation captured by NEMSIS that may indicate EM are clinician impressions, cause of injury, and physical findings. The NEMSIS database records International Classification of Diseases (ICD) codes for each. For 2018, these codes were from the ICD 10^th Revision. The database includes a single primary impression for clinician impressions and allows for multiple secondary impressions. Cause of injury and physical findings also allow for multiple codes to be included. Typically, the codes are generated after the fact by a trained coding specialist based on the notes from the PHC.

The primary outcome of interest was ICD-10 codes related to EM. Specific codes are included in Appendix 1 (Supplementary material). To identify EM, including sub-types, we used a set of ICD-10 codes that have been previously used in research (15). We included each EMS encounter where EM was documented via ICD-10 code in any category of clinician impressions, cause of injury, or physical findings for analysis. These codes allow for nonspecific abuse determinations, such as T74.9—“Unspecified maltreatment—confirmed.” Additionally, we extracted and examined demographic data, injury location, and associated physical exam findings (comprising physical, behavioral, and psychological findings).

Statistical Analysis

First, we described the demographic characteristics of the study population using standard counts, means, and percentages. We explored the prevalence of EM subtypes and associated physical exam findings. For secondary aims, we compared these physical findings across EM subtypes and by anatomic locations. We used R (version 4.1.1) for statistical analysis.

Results

The NEMSIS database included 9,605,522 EMS encounters for patients aged ≥60 during the study period. In 1,765 encounters (0.02%), coders assigned ICD-10 codes consistent with EM based on PHC documentation. Most of these were listed as the cause of injury (64%), followed by PHCs first impressions (25.4%). Elder mistreatment was indicated in the clinician record in more than one location 125 times (7%). The median age of patients with documented EM was 72 (IQR = 65–82), and 62.3% were female. Of those with race reported (n = 484, 27.4% of all encounters with PHC documentation of EM), 64.8% were White (see Table 1 for demographics). Trends were similar across abuse subtypes (see Table 2), with female gender and white being the most common categories. Interestingly, the average age for physical abuse victims was 5–6 years younger than those from other abuse types. The most reported location for service was at a residence. However, 212 (12%) were reported at nursing homes, well above the national average of 1–2% of older adults (16). Of note, there were 24,952 dispatch complaints of assault. Only 408 (1.64%) of these were coded as EM. If all dispatch codes for assault were coded as EM, that would increase the prevalence from 0.02% to 0.26%, some 13 times higher rate. These cases were most coded as T14.90 Injury, unspecified, S09.90 Unspecified injury of the head, and G89.1 Acute pain, not elsewhere classified.

Table 1.

Identified elder mistreatment in NEMSIS 2018 sample descriptive statistics.

Variable	N (1,765)	%
Age (Med./IQR)	72	65–82
Gender
Female	1,100	62.32
Male	652	36.94
Race (N= 486)
White	315	64.81
American Indian or Alaskan Native	15	3.09
Asian	9	1.85
Black or African American	110	24.49
Hispanic or Latino	36	7.41
Native Hawaiian or Pacific Islander	1	0.21
Missing (not included in percentages)	1,279
Disposition
Treated and Transferred by this EMS	1,254	71.05
Released AMA	150	8.50
Refused Treatment	88	4.99
No Treatment	47	2.66
Dead at Scene	6	0.34
Other	220	12.46

N= 1,765 unless otherwise noted. It may not equal 100% due to missing data (coded as “not recorded” or “not applicable”).

Table 2.

Summary of demographics by abuse sub-type.

Race N (%)				Gender N (%)		Age
Abuse Type	White	Black	Hispanic	Other	Male	Female	Mean	SD
Physical Abuse	56 (15.22%)	13 (3.53%)	5 (1.36%)	4 (1.09%)	154 (41.85%)	210 (57.07%)	70.82	10.71
Psychological Abuse	1 (16.67%)	0 (0%)	0 (0%)	0 (0%)	3 (50%)	3 (50%)	76.50	11.24
Neglect	44 (25.73%)	13 (7.6%)	3 (1.75%)	9 (5.26%)	65 (38.01%)	103 (60.23%)	76.69	9.97
Other Abuse	129 (14.02%)	44 (4.78%)	20 (2.17%)	7 (0.76%)	357 (38.8%)	559 (60.76%)	75.83	10.45

N= 1,765. It may not equal 100% due to missing data.

Of the specific types of mistreatments reported, physical abuse was the most common (20.8%), followed by sexual abuse (18.2%), neglect (9.7%), and psychological/emotional abuse (0.34%, see Figure 1). The most common mistreatment type reported was other/non-specified (52.1%). There were 21 encounters with more than one type of EM documented, which corresponds with prior research, which documents that many victims of EM suffer from more than one type of abuse (7). Of these, the most common combination (n = 12) was physical and other abuse. Only one case recorded both physical and sexual abuse. The anatomic location of an injury was provided for 81.3% of all encounters (n = 1,435). Among specific locations, extremity-lower (13.2%), head (10.2%), and extremity-upper (10.0%) were most common. General/global was the most common location descriptor (41.6%). The frequency of anatomic locations of injuries is shown in Figure 2. The most reported physical exam finding was injury or pain (see Table 3). Other commonly reported physical exam findings were psychological, hemorrhage, and cutaneous findings.

Figure 1.

Prevalence of elder mistreatment by subtype from NEMSIS 2018.

Figure 2.

Anatomic location of injury for incidents of elder mistreatment in NEMSIS 2018.

Table 3.

Associated physical, mental, and behavioral findings of elder mistreatment in NEMSIS 2018.

Symptom	N (1,765)	% of all reported symptoms
Pain	469	26.57
Injury	285	16.15
Psychological	178	10.08
Nonspecific	154	8.73
Abuse	117	6.63
Hemorrhage	61	3.46
Cutaneous	43	2.44
Shortness of breath	29	1.64
Nausea/vomiting/diarrhea	18	1.02
Confusion/change in consciousness	17	0.96
Substance use	12	0.68
Fall/difficulty walking	8	0.45
Urinary complaints	7	0.4
Failure to thrive	4	0.23
No Findings	48	2.72
Other	177	10.03

N= 1,765. It may not equal 100% due to missing data.

Discussion

As first responders to medical emergencies, prehospital clinicians have a vital and unique role in identifying and intervening in EM, which is common and severe. For an isolated older adult, an encounter with a PHC may be the only opportunity to recognize mistreatment. Prehospital clinicians are among the few clinicians who evaluate and care for older adults in their homes. Most often, they are mandatory reporters of EM. Despite this, we found PHCs documenting evidence later coded as EM in only 0.02% of encounters. A similar analysis of child abuse documentation also found that it was documented in a small percentage of encounters. Notably, though, child abuse was documented at a rate that was five times higher than EM, despite evidence that EM is a much more common phenomenon (14). The results presented a much higher number of older adults whose primary dispatch complaint was assault. However, only 1.6% of these cases were coded as EM. Approximately 75% of EM was coded without a dispatch code for assault, and 98% of dispatches for assault were not coded as EM. Most often, dispatches for assault were coded as nonspecific injury or pain. This suggests a significant opportunity to change PHC practice.

When the EM subtype was reported, physical abuse was the most common (20.8%), followed closely by sexual abuse (18.2%). Both are less prevalent than other types of EM, with physical abuse occurring in 2.6% and sexual abuse in 0.9%, in comparison to verbal/emotional/psychological abuse in 11.6%, financial exploitation in 6.8%, and neglect in 4.2% worldwide (17). The finding that PHCs document them more commonly is likely because these types of abuse may lead to injuries that are more readily apparent during a physical examination and are perhaps even the catalyst for emergent medical evaluation in some cases, as demonstrated by the high identification of EM as the cause of injury or clinician first impression. Sexual abuse was documented particularly frequently, in juxtaposition to prevalence studies that suggest this is much less commonly reported than other types of EM (17). This may be because, when concern for sexual abuse existed, it was the primary reason for the activation of the Emergency Response system and the subsequent reporting of this concern to PHCs when they arrived on the scene.

That sexual abuse was the primary impression in 43% of encounters in which it was documented supports this. This finding that physical abuse and, particularly, sexual abuse are the specific types more commonly documented by PHCs is like findings for child maltreatment (14). Notably, though more common and associated with particularly high morbidity and mortality (4), neglect was uncommonly documented by PHCs. This is also consistent with findings in child maltreatment. In examining child maltreatment, the authors stated that PHCs may not feel comfortable reporting neglect based on a single encounter without a longer-term understanding of the care being provided, requiring a more longitudinal perspective to report neglect (14). This may be similarly true for EM. Yet, given that PHCs have a unique opportunity to assess the safety of the home environment, increasing recognition, documentation, and reporting of elder neglect is critical to improving these older adults’ health and quality of life. Psychological/emotional abuse was also very infrequently documented, likely because there are seldom immediately apparent medical consequences from this. Notably, when EM was documented, coders often chose a general rather than a specific diagnosis, selecting “other/unspecified” instead of something specific, such as physical abuse. This may represent a lack of confidence in detection of the phenomenon as well as concern about being specific in documenting/diagnosing a specific type of EM in the context of a short encounter (9). Consistent with prior research (3). We found more cases in nursing homes, suggesting that these populations are at heightened risk for EM.

Head and extremities were the most common in encounters with the specific location of injury documented. This is consistent with previous investigations of injuries in physical elder abuse (18). Pain and injury were the most associated symptoms. This likely results from the large percentage of encounters that involved physical and sexual abuse.

Little research has examined the identification of EM in prehospital care (9, 19). Prehospital clinicians reported thinking the issue was familiar but did not commonly identify or report it (20). In one study, 68% of clinicians reported suspecting EM during the last year (mean, 2.3 cases per year), but only 27% of these cases (mean = 0.62 cases per year) were reported. Reasons for not reporting included (1) lack of knowledge about what constitutes EM, (2) lack of awareness of mandatory reporting laws, (3) no understanding of the process for reporting and the appropriate agency to which to report, and (4) confidentiality concerns (21). A recent follow-up study in the same population, conducted 25 years after the above research, found a slightly higher reporting rate, but the reasons for not reporting had not changed (22). Other work has described additional barriers for PHCs to take the next steps when concerned about EM, including challenges communicating concerns to Emergency Department (ED) clinicians, time limitations, lack of protocols, and lack of feedback on the results of their reports (9). While many PHCs report feeling comfortable caring for older adults, research has shown important training deficits exist, including in age-specific medical issues, including dementia, polypharmacy, psychosocial issues, access to social services, and advanced care planning (23). This previous research provides context and a possible explanation for our finding that PHCs very seldom document EM.

Moreover, based on the acuity of the situation, PHCs may only do a limited amount of assessment for medical and social issues other than the primary complaint. In many cases, however, where immediate acute intervention and transport are unnecessary, PHCs typically perform a more comprehensive medical and social history and observe the patients’ living circumstances and interactions with others to form their assessments. This work is highly protocolized and involves the assessment of many social and environmental concerns, such as reporting if the utilities are working, if there are insect infestations, and, in the case of minors, if child abuse is suspected. Given that assessment for child abuse is already predominantly occurring (24), it is not unreasonable to expand the scope of work to include other at-risk populations, such as older adults. Indeed, qualitative studies with PHCs show they are willing and interested in detecting and reporting EM (9).

Several potential solutions exist to overcome these barriers and improve the recognition and reporting of EM among PHCs. Additional education about EM should be integrated into PHC training and continuing education. Prehospital clinicians have reported an interest in further education on this topic (9). Education Development Center, Inc. recently developed an interactive online course funded by the RRF Foundation for Aging. The course Recognizing and Responding to Elder Mistreatment is available through the National Association for EMTs (NAEMT) and approved by the Commission on Accreditation for Prehospital Continuing Education (CAPCE). The course includes four modules: (1) Introduction to Elder Mistreatment and the Role of EMS, (2) Trauma Informed Care with Older Adults, (3) Recognizing Warning Signs of Elder Mistreatment, and (4) Responding to Elder Mistreatment (25).

Additionally, the care that PHCs provide is highly protocolized (13). Therefore, incorporating EM into these protocols can potentially codify a new standard of PHC practice. Unfortunately, an examination of state-level EMS protocols revealed that only 40% contained any mention of EM, with an even smaller percentage providing specifics related to identifying or reporting. Focusing on developing comprehensive protocols and integrating them into state- and agency-level guidelines offers an opportunity to impact the practice of many PHCs (13). Most existing PHC protocols focus on physical indicators of EM, such as fractures and burns. New protocols should also describe signs suggestive of neglect, particularly given that PHCs’ access to the home offers an opportunity to observe safety issues (e.g., lack of heat, vermin infestation, no food in the refrigerator) that other healthcare clinicians may not have. Additionally, these protocols should optimize care and safety for those who reject transport (9). This may include referrals to local food pantries or relevant social services programs (13).

Standardized screening and assessment tools for EM designed for PHCs may also improve identification and documentation. The DETECT tool is a promising checklist instrument that has been developed and tested (6). Integrating a tool such as DETECT into the PHCs’ electronic record can ensure that they consider and assess for EM and document any findings in a standard way. Optimally, the PHC electronic record should include specific fields to enter information about reporting EM to the authorities. This should minimize the time needed to accomplish this task, as PHCs are evaluated on the efficiency of their calls/turnaround times.

Limitations

Using the NEMSIS database to explore EM has limitations. Though it includes most EMS encounters, it is a convenience sample dependent on agency reporting. Agencies with more resources may be more likely to report to NEMSIS than those with fewer resources. Additionally, our approach relies on diagnostic documentation to indicate EM detection by PHCs. It is possible that, in some encounters, PHCs may have recognized EM and taken additional steps (e.g., discussing with ED medical clinicians or social workers) without documenting it. Further, different EMS agencies may have varied protocols for appropriate documentation of EM and other family violence. This database does not include information about whether a report to Adult Protective Services or law enforcement was made, so we cannot determine how frequently these reports were made by PHCs for encounters with EM documented or for other encounters in which it was not. This dataset is not linked to ED or hospital data. Therefore, the final disposition of EMS encounters is unknown, including whether the ED/hospital was aware of PHC concerns and, if so, what next steps were taken, if any. Also, short- and long-term outcomes for these older adults are unavailable. Finally, NEMSIS collects data on EMS encounters, not individuals, so that some people may be in the dataset multiple times. Still, it is impossible to identify such cases in the data. Despite these limitations, NEMSIS has been used extensively and remains an excellent data source for exploring this important research question. Future work should further examine factors that contribute to the under-recognition of EM and continue to develop and test interventions to improve the detection and reporting of EM among PHCs.

Conclusions

Elder mistreatment is a serious, growing public health problem, and PHCs are ideally situated to contribute to increased identification, reporting, and initiation of intervention. Despite this, PHCs very seldom recognize and document this in the older adults they care for. As with child maltreatment, EM is most likely to be identified and documented when there are overt signs of trauma or specific concerns about sexual abuse. Additionally, comprehensive training and new protocols are needed to ensure better identification and reporting of other types of EM, including elder neglect. Empowering PHCs has the potential to have a significant impact.