Gender Differences in Patient-described Pain, Stress, and Anxiety Among Patients Undergoing Treatment for Painful Conditions in the Emergency Department

Abstract

The primary objective of this study was to determine whether patient-described pain correlates with patient-described stress, anxiety, and satisfaction with ongoing treatment and if that correlation differs by gender. This was a prospective observational study at an urban, Level I trauma center conducted between June 1, 2010, and January 1, 2013. Patients reporting pain rated greater than 3 of 10 were eligible. Patients who qualified and consented for participation completed demographic and pain, anxiety, stress, and satisfaction scales at baseline, every 30 minutes, and at discharge. Data were analyzed using analysis of variance, chi-square tests, t-tests, multiple regression, and the Wilcoxon-Mann-Whitney rank test. A total of 7,124 patients were screened for enrollment. Of those, 3,495 (49%) did not qualify at screening for various reasons, including insufficient pain levels (17.5%), elected not to participate (37.7%), did not qualify for other reasons (12.4%), and reason not captured (32.4%). A total of 3,629 (51%) screened patients were eligible and consented. Of those, 620 (16.8%) did not have any data collected past baseline, leaving 3,009 as the final sample size. The patients completing data collection had a median age of 39 years (range = 18 to 90 years), and 50% were male. The mean presenting pain visual analog scale (VAS) score was 71.5 mm. Presenting stress and anxiety VAS scores were significantly higher in females (0.61 and 0.53, respectively) than males (0.56 and 0.50, respectively), whereas presenting pain VAS (0.71 male and 0.72 female) and satisfaction VAS (0.34 male and 0.35 female) did not differ by sex. Ethnicity, education, and income were all statistically different when compared with baseline pain, stress, anxiety, and satisfaction. Male gender was associated with a significant change in pain over time from baseline (coefficient = 0.040, p = 0.037); however, when adjusting for age, ethnicity, education, and income, and for changes in stress, anxiety, and satisfaction VAS scores, changes in pain related to male gender was no longer significant (coefficient = 0.034, p = 0.11). When asked about their satisfaction with the results of the pain treatment that had been provided, patients reported a median of 2 (out of 6, 1 = satisfied, 6 = dissatisfied; interquartile range = 1 to 2). There was no significant difference between sexes (p > 0.90). Patient-reported stress and anxiety were higher among female patients than male patients, but there was no significant difference in reported pain and satisfaction between sexes. Sex alone was not a significant predictor of change in pain for patients presenting to the emergency department with pain-related complaints. Anxiety and stress may potentially influence the pain–gender relationship. el objetivo principal de este estudio fue determinar si el dolor descrito por el paciente se correlaciona con el estrés, la ansiedad y la satisfacción descrita por el paciente con el tratamiento en curso y si esa correlación difiere en función del género. estudio observacional prospectivo de un centro de traumatología de nivel I urbano llevado a cabo entre el 1 de junio de 2010 y el 1 de enero de 2013. Se eligieron los pacientes que documentaron un dolor clasificado como mayor de 3 sobre 10. Los pacientes que cumplieron los requisitos y consintieron para la participación completaron datos demográficos y de dolor, ansiedad, estrés y escalas de satisfacción basales cada 30 minutos y al alta. Los datos se analizaron utilizando el análisis de la varianza, los test de la ji al cuadrado y de la t de Student, la regresión múltiple y el test de Wilcoxon-Mann-Whitney. Se cribó a 7.124 pacientes para la inclusión. De éstos, 3.495 (49%) no cumplieron los requisitos en el despistaje por varias razones, incluyendo niveles de dolor insuficiente (17,5%), la elección de no participar (37,7%) o que no cumplieron los requisitos por otros motivos (12,4%) y por razones no capturadas (32,4%). Tres mil seiscientos veintinueve pacientes (51%) cribados se eligieron y consintieron. De aquéllos, 620 (16,8%) no tuvieron ningún dato recogido en la visita basal, por lo que el tamaño de la muestra final fue de 3.009. Los pacientes que completaron la colección de datos tenía una mediana de edad de 39 años (rango de 18 a 90 años), y un 50% fueron varones. La puntuación media de dolor a la llegada según la escala visual analógica (EVA) fue de 71,5 mm. Las puntuaciones de EVA de ansiedad y estrés a la llegada fueron mayores en las mujeres (0,61 y 0,53, respectivamente) que en los varones (0,56 y 0,50, respectivamente), mientras que la EVA de dolor a la llegada (0,71 hombre y 0,72 mujer) y la EVA de satisfacción (0,34 hombre y 0,35 mujer) no difirió por el género. La raza, el nivel educativo y los ingresos fueron diferentes de forma significativa cuando se comparó en la visita inicial el dolor, el estrés, la ansiedad y la satisfacción. El sexo masculino se asoció con un cambio significativo en el dolor a lo largo del tiempo desde la visita inicial (coeficiente: 0,040, valor de p: 0,037); sin embargo cuando se ajustó por la edad, la raza, el nivel educativo, los ingresos y por los cambios en las puntuaciones de EVA en el estrés, la ansiedad y la satisfacción, los cambios en el dolor en relación al sexo masculino ya no fueron significativos (coeficiente: 0,034, valor de p: 0,11). Cuando se preguntó sobre su satisfacción con el resultado del tratamiento del dolor administrado, los pacientes documentaron una mediana de 2 sobre un máximo de 6 (1 = satisfecho, 6 = no satisfecho) (rango intercuartílico: 1 a 2). No hubo diferencias significativas entre los géneros (valor de p > 0,90). El estrés y la ansiedad documentada por el paciente fue mayor entre las pacientes mujeres que en los varones, pero no hubo diferencias significativas en el dolor y la satisfacción documentada entre géneros. Únicamente el género no fue un factor predictivo significativo de cambio en el dolor para los pacientes atendidos en el SU con un motivo de consulta relacionado con el dolor. La ansiedad y el estrés pueden potencialmente influir en la relación entre dolor y género. Pain is the motivating factor for many patients seeking emergency care, and as a presenting complaint, pain accounts for up to 70% of visits to the emergency department (ED).1, 2 Despite the fact that so many patients visit the ED for pain, pain is often poorly or undertreated,3, 4 One of the primary reasons for this is that the assessment of pain and clinically significant changes in pain is complex, and not all of the factors involved, both in terms of the patient and the assessor, are fully understood.5, 6 Acute pain is subjectively assessed, and there is not yet a way to objectively measure the amount of suffering a patient is experiencing from pain. Because the measurement of pain depends on the subjective assessment by the provider and the subjective expression of symptoms of the patient, it is especially sensitive to a variety of confounding factors, which have limited the precision and accuracy of measurement devices in research. Gender differences have been identified in epidemiological studies of patients with pain and in clinical pain trials.7-9 It has been reported that women have different pain thresholds and pain tolerances than men.10, 11 Some studies suggest that female patients report higher levels of pain, receive more analgesics, and are less likely to have resolution of their pain.8, 12-14 However, others have found no differences in presenting pain levels or success of pain treatment between genders.14-16 These differences, if they are indeed present, could be due to differences in a person's pain detection or pain expression systems, but could also be due to the accuracy of pain assessment and measurement in research models. A patient's report and a provider's assessment of pain are subject to a variety of factors that are not independently measured when pain is assessed. The anxiety-provoking circumstances of being in an ED may add to the patient's pain perception or the expression of his or her pain and thus confound the measurement of the patient's pain. The extent to which factors related to pain, such as perceived anxiety and stress, affect pain perception and relief is not known. Patient-reported pain, stress, and anxiety measures have been found to be interrelated.17 In a previous study, we found that patient-reported anxiety was associated with reported pain, while a patient's reported stress level was not.18 Additionally, patient satisfaction and pain levels have been shown to be unrelated.19, 20 If patient-reported stress or anxiety are associated with gender differently than they are with the report of pain, they may represent an identifiable confounder in treatment that could be used to improve the accuracy of pain measurements. In this study, we sought to determine if a patient's report of anxiety, stress, or satisfaction with treatment were associated with his or her report of pain and if any of these factors differed by gender. Our goal was to determine if any of these factors are related to pain and can serve as an additional measure to improve the accuracy of the measurement of a patient's pain experience. If any of these factors differ in their relation to pain by gender, it may indicate that the measurement of this factor(s) could help account for gender-related measurement differences. This was a prospective observational study at an urban, Level I trauma center with 98,000 annual ED visits conducted between June 1, 2010, and January 1, 2013. The study was registered with ClinicalTrials.gov, ID NCT01723137. The Human Subjects Research Committee approved this study prior to implementation. Patients presenting to the ED were prospectively screened for study enrollment. Patients who qualified and consented for participation completed demographic questions and pain, anxiety, stress, and satisfaction scales at baseline, at every 30 minutes, and at discharge. Patients reporting pain from any cause that rated greater than 3 out of 10 on a standard verbal pain reporting scale were screened for eligibility. Exclusion criteria included age younger than 18 years; a decreased level of consciousness; an inability or unwillingness to answer questions due to being in too much pain, critically ill, or disinterested; inability to provide informed consent due to language barriers, intoxication, or altered mental status; and patients who had been sexually assaulted or who were being held in police custody for any reason. Treating clinicians assisted the research associates (RAs) in determining patients' clinical eligibility and decision-making capacity. Trained RAs screened all patients presenting to the ED for possible enrollment based on presenting complaints. Patients who appeared eligible were approached and screened for their pain levels using the standard 1 to 10 verbal pain reporting scale. Patients with pain levels greater than 3 and no exclusions were asked to participate, and if they agreed, they provided prospective informed consent. The RAs conducted a brief survey of patient demographics and characteristics. Consenting patients completed 100-mm visual analog scales (VAS) describing their perceived pain, stress, anxiety, and satisfaction at baseline and then every 30 minutes until discharge. The pain VAS was bounded by the words “no pain” and “most pain possible” on either side of a 100-mm line, the anxiety VAS with the words “not anxious” and “completely anxious,” the stress VAS with the words “no stress” and “most stress possible,” and the satisfaction VAS with the words “completely satisfied” and “not satisfied at all.” VAS were chosen due to their rapid and easy administration. Pain, anxiety, stress, and satisfaction VAS have demonstrated utility in studying acute pain and generalized anxiety, discriminating differences in stress levels between groups, and validated satisfaction ratings.5, 21-25 The patients also completed discharge questions at the end of their visits, which captured information about their impressions of their interactions with the providers and their overall satisfaction with their pain treatment. These ratings were ranked on a six-point scale, with 1 being “very satisfied,” 2 being “satisfied,” 3 “slightly satisfied,” 4 “slightly dissatisfied,” 5 “dissatisfied,” and 6 “very dissatisfied.” The primary outcome measurement was differences in pain, stress, anxiety, and satisfaction by patient sex. The secondary outcome measurement was the effect of patient characteristics and stress, anxiety, and satisfaction on changes in pain reported between baseline and the end of the study protocol. Data analysis was conducted using STATA 12.0. Given the minimal literature published regarding gender differences in the variables examined, no prospective power analysis or sample size calculation was completed. However, post hoc power calculation demonstrated adequate power to assess for differences in the change in pain between sexes, along with differences in stress, anxiety, and satisfaction levels. Although VAS data are technically ordinal in nature, we treated these outcomes as continuous to use parametric methods, given the limitations of nonparametric statistical modeling. Normal distribution and homogeneity of variance was assessed before data were analyzed. To assess for differences in presenting pain, stress, anxiety, and satisfaction, VAS scores by age, sex, ethnicity, education, and income were compared using correlation, separate one-way between groups analysis of variance, chi-square tests, or t-tests as appropriate. The effects of age; sex; education; income; ethnicity; and changes in stress, anxiety, and satisfaction VAS scales on amount of pain improvement were assessed using multiple regression. The regression model was built using a direct approach, with the main outcome variable of change in pain, which was treated as a continuous variable. Four data points were found to be influential in the model and were excluded in the regression. The variables in the model did not demonstrate high colinearity. Satisfaction with pain treatment, which was assessed at discharge, was examined using the Wilcoxon-Mann-Whitney rank test. For all analyses, p ≤ 0.05 denotes statistical significance, with no adjustment for the multiple comparisons. In this study, 7,124 patients were screened and approached for enrollment, and 3,495 (49%) were not included for various reasons; 610 (17.5%) had insufficient pain levels (i.e., pain rated at less than 3/10), 1,318 (37.7%) refused (n = 275 of those refused because they were in too much pain), 435 (12.4%) did not qualify for other reasons (did not speak English, vulnerable, previously enrolled, being discharged), and 1,132 (32.4%) were not enrolled for other reasons that were not captured by the RAs. Of those screened, 3,629 (51%) patients were eligible and consented to participate. Of those, 620 (16.8%) did not have any data collected past the baseline measurement due to various reasons including patient discharge and patient refusal, leaving 3,009 patients included in the final data analysis. The patients completing data collection had a median age of 39 years (range = 18 to 90 years) and 50% were male. Those patients included in the analysis versus those who did not continue were of similar ages (p = 0.87) and ethnicity (p = 0.41). However, a significantly greater number of women completed the study compared to men; patients who dropped out prior to study completion were 57% male, compared with 43% female dropout (p = 0.001). Thus, males were more likely to not complete data collection. The mean (±SD) presenting pain VAS was 71.5 (±0.2) mm (Table 1). Presenting stress and anxiety VAS scores were significantly higher in females (0.61 and 0.53, respectively) than males (0.56 and 0.50), whereas presenting pain VAS (0.71 male and 0.72 female) and satisfaction VAS (0.34 male and 0.35 female) did not differ by sex. Ethnicity, education, and income were all statistically different when compared with sex and baseline pain, stress, anxiety, and satisfaction (Table 1 and Table 2). Mean presenting pain and frequency by pain are shown in Table pain and were higher among whereas pain was higher among 1 the pain by sex the ED The median time discharge (i.e., the final were was at the time (i.e., the with a range from the time to the time in pain was as the change in pain from baseline to discharge. The change in pain among females was and the change in pain among males was (p = The change in stress was for females and for males (p = The change in anxiety was and for females and (p = The change in satisfaction was and for females and (p = The variables in the regression model did not demonstrate high and had an of (p 0.001). Male gender was associated with a significant change in pain over time from baseline (coefficient = 0.040, p = 0.037); however, when adjusting for age, ethnicity, education, income, and changes in stress, anxiety, and satisfaction VAS scores, changes in pain related to male gender were no longer significant (coefficient = 0.034, p = Table in stress and anxiety were statistically significant in both the and the models. To that the between changes in stress or anxiety and change in pain were not by gender, regression were conducted including all of the predictor In both of these change in stress and change in anxiety the significant variable (p data not When asked their satisfaction with the results of their ED pain treatment, patients reported a median of 2 range = 1 to 2). There was no significant difference between sexes (p > 0.90). satisfaction with the way ED to their pain had a mean of = 1 to 2). There was no significant difference between sexes (p > This study examined the of patient-described stress, anxiety, and satisfaction with reported pain more on whether these differed by gender. research has shown a VAS difference in pain greater than mm and difference in satisfaction between and 11 mm to be the for minimal clinical Anxiety differences between 10 and mm have been shown in research to be clinically this research was not conducted in the ED No data on thresholds for stress were found in the baseline there were sex differences in reported stress and anxiety, but not in reported pain or satisfaction. The difference in anxiety was the for clinical in Given that no such for stress it is if the difference in stress at baseline among sexes is clinically Additionally, baseline pain, stress, anxiety, and satisfaction differed by ethnicity, education, and sex was not significantly associated with a change in pain from baseline to discharge when other were which with the difference in change in pain between the sexes being the for clinical mm in this The difference in anxiety change was greater than which is at the for clinical of 10 to mm previous The difference in change in satisfaction was also at the to for clinical mm in this satisfaction with the results of their pain treatment did not differ by The differences in changes in stress and anxiety as the factors that were associated with the change in pain differ from previous research that found anxiety but not stress to be associated with These previous that patient-reported assessment of pain not differ by However, the difference in baseline stress and anxiety between sexes as other patient points to these variables as pain assessment and measurement between sexes. The effects of these which have not been measured in acute pain may account for results from research in to study, found sex differences in pain and satisfaction with pain treatment in the If there is a difference in anxiety between sexes, such as in this study, the between pain and anxiety may be a factor that to a measured difference in pain between sexes. that there is a difference in reported anxiety and stress by sex and that a perceived change in anxiety and perceived change in stress are both associated with a perceived change in pain it is possible that the measurement of perceived anxiety and perceived stress could be a of differences in to treatment that is not for by pain pain research for anxiety and stress to the effect of anxiety and stress on the pain–gender relationship. research to the of pain, stress, and anxiety as they to gender may to measurement that account for differences in a patient's perception and expression of pain. This was an observational study with outcomes conducted at This the of the that we can as as the of the In this was a sample of The patients who enrolled in the study may have differed from those who those who did not complete data and other groups excluded may have had different pain than those included in the study and may have the VAS assessment scales used for pain and satisfaction have been validated for use in the stress and anxiety VAS scales have not been validated for use in this These scales were chosen for of in the ED but the of is a and be in research. The for this study have to survey RAs to this by the of the study at the time of however, may have over This may have to data points and possible in VAS scores reported by Additionally, it was not possible to clinicians due to the of the variables assessed, which could have to treatment The between the outcome variables we examined could be by the given to patients while in the which were not for in this this study also did not the which may have a different pain–gender relationship. Patient-reported stress and anxiety were higher among female patients, but there was no significant difference in reported pain and satisfaction between sexes. There were differences in baseline pain, stress, anxiety, or satisfaction by patient characteristics. Gender alone was not a significant predictor of change in pain for patients presenting to the ED with pain-related complaints. Anxiety and stress may potentially influence the pain–gender relationship.