Prehospital acute traumatic pain assessment and management practices in the Western Cape, South Africa: a retrospective review

Of the 2401 records reviewed, 272 (11.3%) patients were ≤ 14 years of age (M = 7, IQR = 3–11) while the remaining 2129 (88.7%) were > 14 years (M = 31, IQR = 24-41) of which 80.5% (n = 1713) were between 15–44 years of age. High acuity patients (SATS red or orange) accounted for 35.0% (n = 839) of all cases (Table 1).

Assault, transport-related incidents and accidental injuries were the three most common types of incidents (See Additional file 1). The specific injuries sustained were noted in the working diagnosis of 1278 (53.2%) patients with 139 (10.9%) of these sustaining more than one injury (Table 2).

Almost 15% (n = 189) of patients reportedly sustained fractures/dislocations/deformities (with 9 (4.8%) sustaining more than one fracture). The most common injury site was lower extremities (n = 78, 41.3%) followed by upper extremities (n = 63, 33.3%) while 5 (2.6%) of these patients injured both upper and lower limbs, and 15 (8%) patients were thought to have a pelvis/hip fracture.

Of the 59 (4.6%) patients who had burns documented, 25 (42.4%) had a percentage of burn area recorded (range 1–80%). The remaining patients either had no description of the burn or had the burn described in terms of location, type of burn and/or burn severity.

A total of 435 (18.1%) patients had a pain score recorded. The median pain score was 6 (IQR 4–8). Seventy-two (16.6%) of the patients with a pain score had at least one repeated pain score recorded. Figure 3 illustrates the proportion of records in which a pain score was recorded by gender and age group (adult and paediatric).

No association was found between a pain score being recorded and age group (≤ 14 versus > 14 years) (p = 0.649), gender (p = 0.139) or final triage colour (p = 0.076). The majority (78.6%) of those with a pain score reported moderate-to-severe pain (Fig. 4). A further 194 (8.1%) patients had the presence of pain and/or tenderness reported in the working diagnosis, but no pain score recorded. In total, pain was recorded in 617 (n = 2401, 25.7%) patients. Of note, the records of numerous other patients indicated injuries likely to be painful for which the presence of pain was not recorded.

A range of physical non-pharmacological management approaches were recorded in the ePCR including haemorrhage control (n = 680, 28.3%), application of splints (n = 106, 4.4%) and burn dressings (n=57, 2.4%).

Only 68 (n = 2401, 2.8%) patients received medication with analgesic properties of which 27 (39.7%) had a pain score recorded. IV morphine was administered to 66 (n = 68, 97.0%) patients while one received IV ketamine and one intramuscular (IM) diclofenac, firstly or only. Of all the patients (n = 66) who received IV morphine, 10 (15.2%) received an additional morphine dose (all adults) and 6 (9.1%) received ketamine [IV or IM] (one paediatric) in addition to the initial morphine. None of the patients received inhaled nitrous oxide.

Of the 342 patients with moderate-to-severe pain recorded, only 7.6% (n = 26) received analgesic medication. Fifty-two (78.8%) of the 66 patients who received morphine were male, while 4 (6.1%) were ≤ 14 years. The patients who received ketamine and diclofenac were both > 14 years and male. No association was found between the administration of analgesic medications and gender (p = 0.054) or age group (≤ 14 versus > 14 years) (p = 0.151).

For six (8.8%) of the patients who received analgesic medication, the highest qualification of the crew was documented as BLS (n = 1, 1.5%) or ILS (n = 5, 7.3%) while for the remaining 62 (91.2%) patients it was documented as an ALS level qualification (Emergency Care Technician (ECT): n = 22, 32.3%, ALS: n = 32, 47.1% and Emergency Care Practitioner (ECP): n = 8, 11.8%). An association was found between the documented highest qualification (BLS, ILS and all ALS levels) and the administration of analgesic medication (p < 0.001). Patients, where the highest qualification was documented as an ALS (ECT, ALS and ECP) level, were more likely to receive analgesia medication. The strength of the association was weak (ϕ_c = 0.242).

Additionally, an association was found between a pain score being recorded and the administration of analgesic medication (p < 0.001). Patients with a pain score were more likely to receive analgesic medication; however, the strength of association was weak (ϕ = -0.096). An association was also found between pain severity (mild, moderate or severe pain) and the administration of analgesic medication (p = 0.001). Patients with severe pain were more likely to receive analgesic medication; however, the strength of association was weak (ϕ_c = 0.188).

Our findings indicate that many patients that sustained injuries likely to be painful while only a quarter of patients had pain recorded in some form. Less than a fifth of patients had pain measured with a pain assessment tool with the prevalence of moderate-to-severe pain found to be high (> 75%). International studies, likewise, report the prevalence of pain among trauma patients to be high (> 70%) [8, 31] with a high likelihood of moderate-to-severe pain [5, 9, 11, 12]. This study found pain assessment practices to be poorer than those reported by most international studies [7, 8, 18, 19]. The results, however, are similar (18.1% versus 21%) to pain assessment practices previously found among ALS practitioners in Cape Town, South Africa, with better rates of pain reassessment found in the current study (16.6% versus 6%) [21].

The lack of pain assessment has been identified as a hindrance to adequate pain management [14, 32]. The present study supports these findings as patients with a pain score recorded were more likely to receive medication with analgesic properties. Numerous factors contributing to poor pain assessment documentation have been identified. Being younger, being attended to between 00:00 and 06:00 and shorter transport distances were associated with a reduced likelihood of pain assessment documentation, in children [18]. Adults, in contrast, appear more likely to have pain assessment recorded [33, 34] although this finding is not supported by the current study. Finally, uncooperative patients and communication difficulties have been identified as barriers to pain assessment [18, 35] while the lack of validated age-appropriate pain assessment tools for preverbal children has been identified as a barrier to the management of pain [18].

The use of age-appropriate pain scales as part of general patient care, and regarding all trauma patients with acute pain as candidates for analgesia with regular pain reassessment, is evidence-based recommendations made in a clinical practice guideline (CPG) published in the United States of America [36] and recently adopted for the South African EMS CPGs [37]. Employing observational pain scales is recommended for paediatrics < 4 years [36] and would be more appropriate than the current smiley emoticons found in the ePCR, while the Abbey Pain Scale is a suggested option for the cognitively impaired patient in the prehospital setting [38].

While clear prehospital pain assessment guidelines are helpful, the most frequent reason proposed for the insufficient documentation of pain assessment is a lack of pain knowledge [14, 39, 40]. Several studies have shown that educational activities improve the documentation of pain severity, characteristics and reassessment [41, 42]. In addition to educational activities, EMS systems need to encourage systematic pain assessment and the proper clinical documentation thereof.

Non-pharmacological pain management interventions are more commonly associated with the non-emergency setting. However, cognitive and psychological interventions like reassurance, distraction, and physical interventions like positioning, splinting fractures and burn dressings can all be utilised in the prehospital setting [43]. Psychological interventions sometimes occur inadvertently and are unlikely to be documented in clinical notes. Pain educational initiatives increase awareness and utilisation of non-pharmacological pain interventions in the prehospital setting [41, 42].

Most prehospital studies examine acute pain retrospectively [7, 8, 31] and do not report much, if at all, on non-pharmacological pain management thus limiting comparison. The lack of documentation of non-pharmacological pain management also made an evaluation of these treatments in the current study challenging.

Morphine, ketamine and diclofenac (not in the scope of South African prehospital practitioners)¹ were the only medications with analgesic properties administered during this study. Despite the high prevalence of moderate-to-severe pain, less than 8% of those patients, and less than 3% of all the trauma patients received any analgesia. These results are substantially worse than those reported by studies conducted in high-income countries (HIC) although these also reported prehospital pain relief (any aetiologies) to be poor (8 to 42%) [8, 18, 19].

This study revealed that inhaled nitrous oxide was not used. Similar observations were identified by Matthews et al. [21] in the same setting. In the WC, for most emergency care providers (± 84% are operational BLS/ILS) [24], inhaled nitrous oxide is the only prehospital analgesic option. Pain management decision-making for these practitioners is thus limited to requesting the assistance of a higher qualified practitioner (frequent unavailability), non-pharmacological pain management and transportation to a medical facility for further management. The lack of availability of this treatment is a major barrier to effective pain management. Practitioners must have access to analgesic medications as pain care is both a measure of quality emergency care and a human right [44].

Our findings do not suggest disparity between adult and paediatric pain management. This is not consistent with other studies which suggest that adults are more likely to receive opioid analgesics [32, 34, 45]. Further, studies report that women, regardless of age or pain severity, are less likely to receive analgesia [45‐48], a finding which was not supported in this study. Our findings suggest, like other evidence, that patients with more severe pain recorded are more likely to receive analgesia [46, 49].

A finding which is difficult to explain was that six patients received medication with analgesic properties from crews not licensed to administer those medications. We attribute this to either documentation errors or in some high workload situations emergency care providers may be transporting patients after an analgesic medication had been administered by a higher qualified practitioner on scene.

Findings of the study are concerning; however, consideration must be given to the possible reasons for the apparent poor pain assessment and management practices among emergency care providers in the Western Cape, South Africa.

Studies conducted in HIC identified several constraints to prehospital pain assessment and management [32, 35, 40, 50‐52]. A barrier commonly highlighted is knowledge deficit, attributed to limited attention to pain assessment and management during initial training and a lack of ongoing education [32, 40, 51, 52]. The lack of alternative routes of medication administration, guideline restrictions or inadequacies, the need to obtain permission and the reluctance of medical control to approve prehospital analgesia administration are, also, previously identified constraints [15, 32, 35, 40, 52]. Further barriers include negative feedback from ED staff or supervisors, organisational culture, scarcity of higher qualified practitioner, absence of monitoring guideline adherence and communication [35, 40, 50, 52].

The prehospital setting is a challenging and dangerous work environment with emergency care providers in South Africa increasingly confronted by the threat of violence. In addition to the concerns for personal safety, high workload and demands on emergency care providers, analgesic agents are only available to a small proportion of prehospital practitioners in South Africa. Research to describe barriers and enablers to prehospital pain assessment and management in the South Africa setting may identify further issues. Epidemiological studies, further investigating inequalities in pain assessment and management, as well as the prevalence, assessment and management of pain in medical and obstetric cases, will add to the knowledge base.

Like most other observational studies, retrospective reviews have various potential sources of bias including selection and information bias, uncertainty about generalisability and issues with missing data [53]. Probability sampling strategy was used to minimise sampling bias and select a representative sample of the population to allow generalisability of findings [54]. The random selection of the sample from a broader trauma population of the Western Cape, the high burden of trauma and the profile of EMS in the rest of South Africa, mean that the study findings are likely generalisable to prehospital trauma patients in the rest of South Africa. The results may be less generalisable outside South Africa where the burden of trauma may be different, and the profile of EMS systems differ [55]. International studies suggest medical and gynaecological or obstetric conditions to be less common aetiologies of prehospital acute pain [4, 7], and this is another area for further research in low and middle-income settings. Inaccuracies and poor quality of ePCR clinical notes were the foremost limitations to the study findings; however, this may not be a reflection of clinical practice, as an inherent restriction of the retrospective review methodology is the assumption that if it was not documented, it was not done.