The relationship between patient factors and the refusal of analgesics in adult Emergency Department patients with extremity injuries, a case-control study

1 Introduction

Patients with extremity injuries are frequently treated in the Emergency Department. Examples of these injuries are fractures, contusions or strains and sprains. As these injuries are often painful, treatment usually consists of pain treatment with orally or intravenously administered analgesics, besides injury-specific treatment, such as immobilization or elevation. Although the prevalence of pain in patients with traumatic injuries in the Emergency Department is high, previous studies have consistently shown that pain is treated inadequately [1], [2], [3]. Several factors could contribute to this phenomenon of “oligo-analgesia” in the Emergency Department and have previously been described among health care staff [4]. Lack of assessment or underestimation of pain severity play a role, as pain scores are used infrequently and health care staff tends to underestimate pain severity in general [2], [3], [5], [6]. Moreover, concerns about masking occult injury and fear of side effects related to administration of analgesics might explain inappropriate use of pain medication [7]. However, besides these aspects, patient related factors might be as important as well, as patients might decline analgesics for various reasons. For instance, more than 10% of patients with long bone fractures declined pain medication, despite having pain [8]. Reasons for declining analgesics were not reported in this study. The authors of a 2008 study found that nearly 50% in a cohort of Emergency Department patients with traumatic as well as non-traumatic painful conditions refused analgesics [9]. As expected, patients wanting analgesics had higher pain scores than patients who refused analgesics and the main reason not to want analgesics was that pain was considered tolerable. Other reasons patients reported were that they had already taken analgesics at home and that they wanted to stay alert. Additional reasoning have been published elsewhere and included the belief that analgesics should be given only when pain was unbearable [5]. Moreover, in another study in 203 Emergency Department patients presenting with painful injuries, fear of addiction was the most reported reason for analgesic refusal [10].

As mentioned, pain severity seems of importance in the patient’s decision whether to accept or decline analgesics, however there are conflicting results as another Emergency Department study found no relationship between pain severity and desire for analgesia [11], [12]. Because of these conflicting results and the fact that previous studies mainly evaluated mixed study populations with both traumatic and non-traumatic conditions, the current study was undertaken. The objective was to evaluate differences between patients presenting with extremity injuries who accepted and patients who refused analgesics. Moreover, patient related factors were evaluated in patients refusing analgesics. The rationale of this study was that knowledge of these factors is potentially useful in optimizing future Emergency Department pain treatment and make steps towards patient centered pain treatment instead of protocol centered pain treatment.

2 Methods

3 Results

3.1 Study characteristics

During the recruitment period of 4 weeks between August 1st and September 1st 2016, a total of 253 patients were potentially eligible for enrolment in the study. A total of 55 patients declined analgesic treatment at triage and were matched to 110 patients who accepted analgesic treatment (Fig. 1).

Fig. 1:

Patient flow chart. A total of 414 patients with acute traumatic extremity injuries presented to the Emergency Department during the study period, of which 253 were eligible for inclusion in the study. From this population 55 cases and 110 controls were selected.

Baseline characteristics are shown in Table 1. Most patients were male with an injury of the upper extremity after sustaining blunt direct trauma. There were no statistical differences between the cases and the groups, except for the diagnosis at discharge, as significantly more patients who presented with a wound refused pain medication, compared to patients presenting with fractures or presenting with strains or sprains.

Table 1:

Baseline parameters.

Parameter	Cases (N=55)	Controls (N=110)	Total (N=165)	p-Value
Male sex (%)	39 (70.9)	78 (70.9)	117 (70.9)	>0.99
Median age in years (IQR)	43.6 (30.2–54.6)	38.9 (26.4–51.4)	39.5 (27.1–52.7)	0.29
Injured extremity (%)				0.83
Upper extremity	29 (52.7)	60 (54.5)	89 (53.9)
Lower extremity	26 (47.3)	50 (45.5)	76 (46.1)
Trauma mechanisma (%)				0.17
Direct, blunt	26 (47.3)	66 (60.0)	92 (55.8)
Direct, sharp	12 (21.8)	13 (11.8)	25 (15.2)
Indirect	17 (30.9)	31 (28.2)	48 (29.1)
Time to presentationb (%)				0.72
<6 h	29 (69.0)	67 (65.7)	96 (66.7)
6–12 h	3 (7.1)	12 (11.8)	15 (10.4)
12–18 h	1 (2.4)	6 (5.9)	7 (4.9)
18–24 h	3 (7.1)	4 (3.9)	7 (4.9)
24–48 h	6 (14.3)	13 (12.7)	19 (13.2)
Diagnosis at discharge (%)				<0.01
Sprain or strain	18 (32.7)	55 (50.0)	73 (44.2)
Fracture	11 (20.0)	36 (32.7)	47 (28.5)
Wound	26 (47.3)	19 (17.3)	45 (27.3)

1. ^aIndirect trauma was defined as twisting or rotating trauma mechanism, as well as an inversion of a joint. A direct trauma required a direct force to be applied to the body. ^bIn a total of 21 patients data regarding time to presentation was missing (13 in the cases group and 8 in the control group). IQR=interquartile range; NRS=Numerical Rating Scale.

3.2 Outcomes

NRS pain scores were not distributed normally, as the histogram revealed a distribution that was skewed to the right. The median NRS pain score in the group of patients who declined analgesic treatment was 5.0 (IQR 3.0–8.0) and 8.0 (IQR 6.0–9.0) in the control group who accepted analgesia. The difference in NRS scores between these groups was statistically significant (p<0.01) and the distribution of median NRS pain scores of both groups is shown in Fig. 2.

Fig. 2:

NRS pain scores of cases and controls. Box plots of patients declining analgesics (cases) and patients accepting analgesics (controls), showing distribution of NRS pain scores. The box represents the median NRS value with its 1st and 3rd quartile and the whiskers show the minimum and maximum NRS value in each group. NRS, Numerical Rating Scale.

Of all patients, 29 (17.6%) reported minor pain (NRS pain score 1–3), 39 (23.6%) reported moderate pain (NRS 4–6) and 97 patients (58.8%) reported severe pain (NRS pain score 7 or higher).

Nineteen patients (19.6%) with severe pain, 16 patients (41%) with moderate pain and 20 patients (69.0%) with minor pain did not desire analgesic treatment when offered (p<0.01).

Tables 2 and 3 show the results of the logistic regression models in which refusal of analgesia is utilized as binary outcome. Trauma mechanism, diagnosis at discharge and NRS pain scores entered the multivariate models as independent variables after univariable preselection.

Table 2:

Univariable logistic regression model.

Parameter	OR (95%-CI)	p-Value
Age	1.01 (0.99–1.03)	0.439
Injury localization	1.08 (0.56–2.06)	0.83
Trauma mechanisma		0.17
Indirect	1.39 (0.66–2.93)
Direct, sharp	2.34 (0.95–5.80)
NRS pain score	0.67 (0.57–0.78)	<0.01
Time to presentationb		0.70
6–12 h	0.58 (0.15–2.20)
12–18 h	0.39 (0.044–3.34)
18–24 h	1.73 (0.36–8.26)
24–48 h	1.07 (0.37–3.08)
Discharge diagnosisc		<0.01
Wound	4.18 (1.89–9.27)
Fracture	0.93 (0.40–2.21)

1. All available variables were entered in a univariable preselection model. ^aAs “direct blunt trauma” was the largest category, this was used as the reference category. ^bAs “within 6 h” was the largest category, this was used as the reference category. ^cAs “strains and sprains” was the largest category, this was used as the reference category. OR=odds ratio; 95%-CI=95%-confidence interval; NRS=Numerical Rating Scale.

Table 3:

Multivariable logistic regression model.

Parameter	OR (95%-CI)	p-Value
Trauma mechanisma		0.54
Indirect	1.30 (0.57–2.96)
Direct, sharp	0.67 (0.22–2.03)
NRS pain scoreb	0.64 (0.54–0.77)	<0.01
Discharge diagnosisc		0.29
Wound	2.04 (0.83–5.03)
Fracture	1.12 (0.45–2.76)
NRS pain scoreb	0.70 (0.59–0.83)	<0.01

1. The best performing predicting factors from the univariable regression analysis were entered into a multivariate logistic regression model, in case the p-value was 0.2 or lower. ^aAs “direct blunt trauma” was the largest category, this was used as the reference category. ^bIn order to prevent multicollinearity, NRS pain score was analyzed in one model with trauma mechanism and in another model with discharge diagnosis. ^cAs “strains and sprains” was the largest category, this was used as the reference category. OR=odds ratio; 95%-CI=95%-confidence interval; NRS=Numerical Rating Scale.

In order to prevent multicollinearity, both trauma mechanism and discharge diagnosis were entered into separate models, which were corrected for NRS pain scores. Both indirect and direct sharp trauma mechanism as well as wound and fracture as discharge diagnosis did not significantly or independently predict refusal of analgesics. In both models, NRS pain score at presentation was a significant predictor of refusal of analgesics, as the Odds Ratios were 0.64 (95%-CI 0.54–0.77; p<0.01) and 0.70 (95%-CI 0.59–0.83; p<0.01), respectively. This means that for every increase in NRS by one point, the risk of refusal of pain medication decreases with approximately one third.

The results of the Likert scales of the eight-item Barriers questionnaire are shown in Fig. 3. The three statements patients agreed most with, were: “pain medication should only be used in extreme pain” (45.5% very much agreed and 25.5% agreed); “pain medication decreases the doctor’s ability to appropriately judge the injury” (20% very much agreed and 29.1% agreed) and “people easily get addicted to pain medication” (21.8% very much agreed and 21.8% much agreed). However, regarding the latter statement, a total of 27.3% did not agree at all.

Fig. 3:

Results of the Barriers Questionnaire. The 6-point Likert scale answers to the eight items of the Barriers questionnaire. On the x-axis results are shown in percentages.

The three statements patients did not agree with most profoundly were: “patients are not supposed to talk about their pain” (70.9% do not agree at all and 7.3% do not agree); “pain medication will aggravate my injury” (65.5% do not agree at all and 10.9% do not agree) and “pain medication cannot really control pain” (43.6% do not agree at all and 9.1% do not agree).

4 Discussion

In the current study we evaluated reasons why adult patients presenting to the Emergency Department with extremity injuries decline analgesic treatment. The results show that the lower the median NRS pain score, the more likely it is that patients refuse analgesic treatment. In other words, pain scores were significantly higher in patients who accepted analgesic treatment compared to patients who declined this. The logistic regression model supported this by showing that the initial NRS pain score was inversely related to refusal of pain medication with a mean Odds Ratio of 0.67. In line with these findings most patients who declined analgesia agreed that pain medication should be reserved for extreme pain only.

The authors of an observational, prospective, multicenter study in the United States and Canada in 842 patients with acute pain due to various traumatic and non-traumatic causes reported that 30% of all patients did not want analgesic treatment [12]. Of these, 49% of patients with moderate pain and 16% of patients with severe pain did not desire analgesics. These numbers are comparable to our results, as we found that 19.6% of patients with severe pain and 41% of patients with moderate pain refused analgesics.

It seems that our patients did not find it difficult to talk about pain, as they most profoundly disagreed with the statement that “patients should not talk about their pain”. This is in contradiction to the results of a previous survey performed among New Zealand Emergency Department nurses [18]. These authors identified reluctance of patients to report pain as the most important patient-related barrier to optimal pain treatment. These contrasting results emphasize the importance of communication and explicitly asking patients whether this might play a role in declining analgesia.

A significant number of respondents in the same survey perceived patients’ reluctance to take opioids as barrier as well. It was not clear whether fear of adverse events or fear of dependence played a role. The results from our questionnaire did not explicitly reveal this. Although a total of 40% of our study population agreed or agreed very much with the statement of addiction, 27.3% did not agree at all. Additionally, 36.4% of our patients did not agree with the statement to rather have pain than side effects. The same was found by Singer et al. who reported that a only a minority of 7% of patients who declined analgesic treatment gave as reason that they did not want to become groggy and wanted to stay alert [9].

The same authors reported that the most common patient-related factor in their study was that pain was tolerable. Stalnikowicz et al. described that 66% of all patients who were interviewed regarding misconceptions about pain management agreed with the statement that pain medication should only be used when pain was unbearable [5]. These findings match our study results as one of the three statements patients agreed with most, was that “pain medication should only be used in extreme pain”.

Besides this belief, nearly half of our patients agreed or agreed very much with the statement that pain medication would decrease the ability of the physician to judge the injury. This statement resembles the clinical myth that opioids should not be administered before assessment of a patient with abdominal pain, as this would “cloud” physician’s judgment. This myth was busted by the authors of a Cochrane review and therefore patients should be informed that administration of pain medication does not cloud judgment nor distracts the physician from the injury, as results of this review can probably extrapolated to patients with extremity injuries [19].

From our study results it becomes clear that regarding pain and pain medication, education of patients in the Emergency Department is of vital importance. We should inform them that analgesia is not only available for severe or extreme pain and that treating pain with pain medication does not mask symptoms or cloud judgment. As Ducharme concluded in 2013; in order to identify patient-related factors and beliefs, we have to start listening to our patients [20]. If patients do not want pain medication, we should ask them “why not?”

5 Limitations

We designed this study as a matched case-control study in order to isolate a cohort of patients who declined analgesia, a comparable group who did not decline analgesia and be able to correct for gender as this has been correlated with higher pain scores in previous research. There are several other factors we could also have matched for and by not doing so, we might have introduced bias. Examples are ethnicity and discharge diagnoses, as both might be important variables in analgesic treatment and pain severity.

Although an association between parameters can be found by choosing a case-control study design, a definitive cause-effect relationship cannot be determined [21].

We did not investigate whether patients received analgesics later on during their Emergency Department stay. It could have been that some patients changed their mind or only expressed their desire for pain medication to a physician. More importantly, pain medication already taken at home before Emergency Department presentation, were not accounted for.

In this study we used the Barriers questionnaire to evaluate pain experiences, beliefs and attitudes among patients with acute pain. It must be emphasized that this questionnaire has been validated in chronic pain patients and not in the acute setting. Moreover, as only patients who declined pain medication filled in this questionnaire, no comparison of motivations and beliefs between patients who accepted pain medication and patients who declined pain medication could be made.

The study population consisted of a convenience sample as the research assistant who collected patient data was not present in the Emergency Department all the time. Therefore, selection bias might have played a role, although it must be mentioned that only a minority of potentially eligible patients presenting to the Emergency Department and were not accounted for in retrospect. Moreover, patients were included in a level one trauma centre. Therefore, results might not be representative for all Emergency Departments.

Although the study was powered to detect a difference in NRS pain score between the two groups, the study population was relatively small and might not have been large enough to detect differences in other study variables.

6 Conclusions

In adult patients presenting to the Emergency Department with extremity injuries, pain severity is important in the decision to accept or decline analgesic treatment. However, common patient beliefs such as that analgesics should only be used in extreme pain, use of pain medication might prevent the doctor from appropriately judging the injury and fear of addiction, might also play a role in declining analgesic treatment. Therefore, in order to optimize pain treatment in the Emergency Department, these factors should actively be addressed by the emergency health care provider in case patients refuse analgesic treatment.