ACADEMIC EMERGENCY MEDICINE 2012; 19:1166–1172 © 2012 by the Society for Academic Emergency Medicine
Objectives: Patient and parent satisfaction are important measures of quality of care. Data are lacking regarding satisfaction with emergency procedures, including laceration repair. The objective was to define the elements of care that are important to parents during a pediatric laceration repair and to determine the predictors of excellent parent satisfaction.
Methods: This was a cross-sectional observational study of a convenience sample of patients younger than 18 years of age presenting for laceration repair to an urban tertiary care children’s hospital emergency department (ED). At the end of the ED visit, parents completed a survey developed for this study assessing ratings of their experience and their perception of how their child experienced the repair. Exploratory factor analysis was used to derive the factors comprising parents’ perception of the laceration repair process. A separate factor analysis was performed for the 0- to 4-years age subgroup. Multivariate logistic regression was used to determine which of these factors predicted excellent parent satisfaction with the visit, and also satisfaction with the procedure itself, adjusting for sociodemographic factors.
Results: A total of 408 parents returned completed surveys (response rate = 76%). Factor analysis revealed that three factors provided a summary of the 16 survey items. They were labeled “provider performance,”“anxiety and pain,” and “cosmetic appearance,” based on factor loading patterns. Provider performance was the only predictor of satisfaction with the visit (adjusted odds ratio [OR] = 11.6; 95% confidence interval [CI] = 6.2 to 21.6). Provider performance (adjusted OR = 4.7; 95% CI = 3.1 to 7.2) and cosmetic appearance (adjusted OR 2.7; 95% CI = 1.7 to 4.2) predicted satisfaction with the procedure. Anxiety and pain did not predict either outcome.
Conclusions: Provider performance, which comprises the elements of physician communication, caring attitude, confidence, and hygiene, is the strongest predictor of excellent parent satisfaction for pediatric patients with ED visits for laceration repair.
Objetivos: La satisfacción del paciente y de los padres son importantes medidas de la calidad de la atención. No existen datos sobre la satisfacción con los procedimientos de urgencias, y ello incluye la reparación de una herida. El objetivo fue definir los elementos de la atención que son importantes para los padres durante la reparación de una herida en pediatría y determinar los factores predictivos de una satisfacción excelente de los padres.
Método: Estudio observacional transversal de una muestra de pacientes menores de 18 años de edad que acudieron para la reparación de una herida a un servicio de urgencias (SU) de un hospital infantil terciario y urbano. Al final de la atención en el SU, los padres rellenaron una encuesta diseñada para este estudio que valoraba su experiencia y la percepción de cómo su hijo experimentó la reparación. Se usó un análisis factorial exploratorio para derivar los factores de la percepción de los padres en el proceso de reparación de la herida. Se realizó un análisis factorial por separado para el subgrupo de 0 a 4 años de edad. Se utilizó una regresión logística multivariable para determinar cuáles de estos factores predecían una satisfacción excelente de los padres con la atención, y también la satisfacción con el procedimiento en sí mismo, y se ajustó por los factores sociodemográficos.
Resultados: Cuatrocientos ocho padres devolvieron las encuestas completadas (porcentaje de respuesta del 76%). El análisis de factores reveló que tres de ellos proporcionaban un resumen de los 16 ítems de la encuesta. Se etiquetaron como “Rendimiento del médico”, “Ansiedad y dolor”, y “Apariencia cosmética”, en base a los factores considerados. El “Rendimiento del médico” fue el único factor predictivo de satisfacción con la atención (OR ajustada 11,6; IC95% = 6,2 a 21,6). El “Rendimiento del médico” (OR ajustada 4,7; IC 95% = 3,1 a 7,2) y la “Apariencia cosmética” (OR ajustada 2,7; IC 95% = 1,7 a 4,2) predijeron la satisfacción con el procedimiento, mientras que “Ansiedad y dolor” no predijo ningún resultado.
Conclusiones: El “Rendimiento del médico”, el cual incluye elementos de comunicación, actitud en el cuidado, confianza e higiene, es el mayor predictivo de una satisfacción excelente de los padres de los pacientes pediátricos que son atendidos en el SU para la reparación de una herida.
Patient-centered care is one of the six components of high-quality care as determined by the Institutes of Medicine. It is incumbent on health care providers to determine what patients and parents value in health care.1 There have been several studies examining factors that correlate with satisfaction in the emergency department (ED) setting, and some important associations have been identified. These include robust provider–patient information exchange,2,3 empathetic provider–patient interactions,2,4 and minimization of perceived wait times.5,6 However, studies to date have primarily examined heterogeneous samples of patients with regard to their reason for seeking emergency care. They rarely focus on the important subgroup of patients whose ED experience revolves specifically around invasive procedures. We chose to explore which elements of the pediatric laceration repair process were important to parents and to determine predictors of excellent care related to laceration repair.
There are many components to an ED experience that influence satisfaction with care, including parental anxiety about the need for an emergency visit, unfamiliar ED processes, interactions with a new health care team, and concerns around timeliness. Specifically, laceration repair in children can be an anxiety-provoking experience for children, parents, and physicians. It is our expectation that elucidating the factors comprising parents’ satisfaction with laceration repair could serve as a platform for improving patient care in this specific area, as well as broader aspects of pediatric emergency care.
We used exploratory factor analysis to determine elements of care that influence parental ratings of their child’s experience during laceration repair. Using the factors obtained from the factor analysis, we determined the predictors of excellent parental satisfaction with the visit overall, as well as excellent satisfaction with the procedure itself in a population of pediatric patients presenting to the ED for laceration repair.
Study Design and Population
We performed a cross-sectional observational study of pediatric patients presenting for laceration repair, designed to determine the elements of care that predicted excellent parent satisfaction. The study was granted expedited approval from the hospital’s institutional review board. Verbal consent was obtained from all participants.
The study was conducted at an academic urban tertiary care children’s hospital ED. The ED at this hospital sees approximately 60,000 patients per year and is a major referral center for young children with lacerations and patients with complex lacerations. Subjects consisted of a convenience sample enrolled over a 6-month period from July 2010 to December 2010, 7 days a week from 2 p.m. to midnight. Recruitment was done by trained ED research coordinators and the principal investigator. Patients were eligible for participation if they were younger than 18 years of age and had laceration repair performed in the ED during that visit. We included patients who had procedural sedation, multiple lacerations, and patients who had their repairs performed by plastic surgeons. Patients were excluded if they had multisystem trauma, required hospital admission, or had no accompanying parent to complete the survey or if the parent or guardian was neither English- nor Spanish-speaking.
Survey Content and Administration
The survey was created de novo in collaboration with an expert in survey methodology. The survey items included a combination of previously reported measures with known psychometric properties2,7–9 and newly developed questions for the purposes of the study. There were multiple revisions of the survey. A pilot was performed on a group of parents using cognitive interviewing techniques to assess the ease of completion and clarity, both of which were found to be adequate. The survey was professionally translated into Spanish. Translation back into English by a different translator, and comparison to the original survey, was used to assess the equality of both language versions of the questionnaire.
Patients were identified from the ED electronic tracking board (FirstNet, Cerner Corporation, Kansas City, MO) on arrival to the ED. Parents of eligible patients were approached for enrollment after the wound was repaired and while awaiting formal discharge instructions. Verbal consent was obtained from the parent, and an information sheet about the study was provided. It was emphasized that the treating physician would not see the responses and that the care would not be influenced in any way. The parent(s) privately completed the self-administered paper survey (Data Supplement S1, available as supporting information in the online version of this paper) and placed it into an envelope and into a specified box or returned the envelope directly to study personnel. Parents also had the option of mailing back the completed survey. The survey asked them to rate aspects of their visit and their satisfaction with the care they received. Parental sociodemographic information was also collected including age, relation to patient, education level, race, and ethnicity.
Survey data were entered into a database in REDcap (Vanderbilt University, Nashville, TN) by the research coordinators and primary investigator. REDCap is a secure, Web-based application specifically designed to build and manage online databases for data capture in clinical research. Accuracy of all data entered from the paper surveys into the online database was cross-checked by an independent research coordinator.
Our primary outcomes were overall parental satisfaction with care and satisfaction with the procedure itself. For the purposes of performing a regression analysis, we dichotomized ratings of satisfaction from a five-point Likert scale into “excellent” versus “very good, good, fair and poor.” We used this cut-point because we expected our satisfaction ratings to be highly skewed toward excellent based on prior unpublished hospital satisfaction data, where 72% of parents rated their ED care as excellent.
We subjected the 16 survey items measuring different aspects of parents’ perception of the laceration repair process to an exploratory factor analysis using STATA 11 (2009, StataCorp, College Station, TX). Factor analysis is a statistical method intended to explain the relationships among several difficult to interpret, correlated variables in terms of a few conceptually meaningful, relatively independent factors.10 The factors are not measured directly, but are inferred from the variables that represent the factor and hence they are also referred to as underlying or latent constructs. For example, in the field of psychology, a factor called “mathematical intelligence” might contain variables such as algebra, geometry, and trigonometry. This factor would be relatively distinct from “verbal intelligence” containing variables such as spelling, grammar, and reading comprehension, even though both mathematical and verbal intelligence contribute to overall intelligence. Factor analysis techniques have been used to validate emergency medicine related survey instruments.11
Surveys with any missing responses were regarded as incomplete and were not included in the analysis. Factors were extracted using the principal factor method with the five-point Likert scale in its original format. We determined the number of factors to be retained based on their eigenvalues and by the change in slope of the scree plot. We then performed an oblique rotation to the factor solution, allowing some degree of correlation between the factors. Variables that did not load on any of the rotated factors were removed, and the factor analysis was repeated to produce the final factor solution. We used a recommended factor loading cutoff of 0.3 to determine variable retention.12 We then calculated factor scores for each retained factor. We assessed the relationship between the factor scores and our primary outcomes using Spearman correlation coefficients, using a Bonferroni adjustment to correct the familywise Type 1 error rate (critical value for Spearman tests = 0.008). We estimated a multivariate logistic regression model to predict excellent satisfaction with the visit and excellent satisfaction with the procedure itself, as a function of the factor scores, adjusting for demographic characteristics (patient age, parent age, race, ethnicity, and education level). We evaluated the overall model performance with the likelihood ratio chi-square test. Cronbach’s alpha was used to evaluate the internal consistency of the questionnaire as a whole, and for each of the independent factors.13
Our sample size calculation was based on the generally accepted subject to variable ratio of 10 to 15 subjects for each analyzed variable.12 The parent survey contained 16 items that measured different aspects of the patient’s laceration care encounter, which we subjected to factor analysis. All other survey items were of a demographic nature or related to other experiences. We therefore estimated that we would need 160 to 240 subjects to perform the factor analysis. We collected more than this minimum number to accommodate subgroup analysis in the 0- to 4-years age group. We decided a priori that this group was important to study separately because they are generally the most challenging patients to perform repairs on, often requiring procedural sedation, and anecdotally having parents with high levels of stress and anxiety.
Figure 1 shows the patient enrollment flow chart. The survey response rate was 76%. Mailed surveys accounted for 3.7% of respondents. Patients of respondent parents were similar in age and sex to those who refused to participate (nonrespondents). The mean (±standard deviation [SD]) patient age was 6.0 (±4.3) years for respondents versus 5.9 (±4.8) years for nonrespondents (mean difference = 0.1 years; 95% confidence interval [CI] = −0.9 to 0.8). There were 59% male patients in the respondent group compared to 67% for nonrespondents (risk difference = 8%; 95% CI = −1.6 to 16.7). Tables 1 and 2 show parent characteristics and wound characteristics, respectively. Most respondents were mothers and the majority were white, non-Hispanic, and educated at or above the college level. The majority of parents rated satisfaction with care as either very good or excellent.
|Parent age, median (IQR)||38 (33–42)|
|Relation to patient|
|Black or African American||54 (13.2)|
|Did not specify||16 (3.9)|
|Did not specify||16 (3.9)|
|High school degree or less||53 (13.3)|
|Some college||75 (18.8)|
|College graduate or higher||265 (66.4)|
|Did not specify||6 (1.5)|
|Visit satisfaction ratings|
|Very good||106 (30)|
|Type of closure|
|Local infiltration||216 (61.9)|
|Oral midazolam||85 (24.6)|
|Procedural sedation||35 (10.2)|
|Provider performing repair|
|Plastic surgery||40 (9.8)|
Comparing the 0- to 4-years age subgroup to the remaining cohort, there were significantly higher rates of anxiolysis (34.5% vs. 15%, respectively), procedural sedation (14.6% vs. 6.1%), and facial lacerations (76% vs. 56%). Parent self-ratings of their anxiety, and parent perception of patient anxiety, were also significantly higher in the 0- to 4-years age subgroup. There were no significant differences between the two groups in pain scores or the level of training of the providers who performed the repairs.
A total of 408 patients had complete survey data and were included in the factor analysis. The internal consistency of the survey as assessed using Cronbach’s alpha was 0.84 using all 16 items in the scale and 0.87 for the 12 items used in the factor analysis. From the initial factor analysis, we retained three factors based on a combination of their eigenvalues and the change in the slope of the scree plot (Figure 2). Examination of the factor loadings for the rotated factors revealed that three variables did not load adequately on any of the three factors (patient anxiety before the procedure, parent anxiety during the procedure, and perceived length of stay: the factor loadings for these variables were 0.21, 0.23, and 0.19, respectively). These were removed and the factor analysis was repeated. This second analysis revealed that the largest factor loading for the variable “How much was done to control the patient’s pain” was close to our cutoff for exclusion (0.32). Also, the uniqueness (i.e., the percentage of variance for the variable that is not explained by the factors) of this variable was particularly large, at 86%, suggesting that it was measuring a construct that was not meaningfully related to the three retained factors. These results compelled us to drop this variable and run the factor analysis again to produce the final factor solution.
Based on the pattern of factor loadings of the final factor solution and the content of these variables, we labeled the three retained factors as “provider performance,”“anxiety and pain,” and “cosmetic appearance.” It is important to note that the labels assigned to the factors were descriptive summary terms for the variables comprising them, so for instance, provider performance did not represent technical competence, but rather was a term we created to summarize the six variables comprising that factor. The variables contained in each factor and their respective factor loadings are shown in Table 3. Cronbach’s alpha was 0.90 for the six items in provider performance, 0.85 for the four items in anxiety and pain, and 0.57 for the two items in cosmetic appearance.
|Survey Item||Main Analysis (n = 408)||0- to 4-Years Age Group (n = 194)|
|Provider Performance||Anxiety and Pain||Cosmetic Outcome||Provider Competence||Caring and Kindness||Anxiety and Pain|
|1. Explanation given before procedure||0.57*||−0.03||0.15||0.40*||0.34||−0.03|
|2. Patient anxiety during procedure||0.07||0.71*||−0.02||−0.03||0.02||0.76*|
|3. Patient anxiety compared to expected||0.00||0.79*||−0.02||−0.09||0.01||0.82*|
|4. VAS pain score||0.00||−0.72†||−0.03||0.00||−0.10||−0.71†*|
|5. Pain during procedure||−0.06||0.82*||0.08||0.02||−0.01||0.81*|
|6. Measures to control patient’s pain||N/A||N/A||N/A||0.40*||−0.10||0.10|
|7. Caring attitude of physician||0.86*||0.07||−0.12||0.13||0.82*||0.05|
|8. Caring attitude of nurse||0.74*||0.09||−0.16||−0.01||0.82*||0.10|
|9. Physician communication||0.80*||−0.04||0.12||0.56||0.38||−0.05|
|10. Physician hygiene||0.67*||−0.05||0.22||0.58*||0.35||−0.05|
|11. Physician confidence||0.63*||−0.06||0.35||0.72*||0.19||−0.07|
|12. Cosmetic appearance after repair||0.31||0.02||0.54*||0.75*||−0.03||0.00|
|13. Cosmetic appearance compared to expected||0.02||0.14||0.45*||0.44*||−0.16||0.20|
In univariate analyses using Spearman correlations, provider performance, anxiety and pain, and cosmetic appearance were significantly correlated with satisfaction with the visit (r = 0.61, 0.26, and 0.51, respectively) and satisfaction with the procedure (r = 0.65, 0.30, and 0.41, respectively; all p < 0.008). In the multivariate regression model predicting excellent satisfaction with the procedure, provider performance, and cosmetic appearance were significant independent predictors, with provider performance being the stronger of the two. In the model predicting excellent satisfaction with the visit, only provider performance was a significant predictor. The anxiety and pain factor was not a significant predictor of either outcome (Table 4). Adjusting for patient age, parent age, race, ethnicity, and education level in the regression models did not change the pattern of significance for the factor scores.
|Predictors||Adjusted OR (95% CI)*‡|
|Procedure Satisfaction||Visit Satisfaction|
|Provider performance||4.67 (3.1–7.2)†||11.6 (6.2–21.6)†|
|Anxiety and pain||0.97 (0.7–1.4)||1.35 (0.9–1.9)|
|Cosmetic appearance||2.71 (1.7–4.2)†||1.07 (0.7–1.7)|
We repeated the factor analysis for the subgroup of 194 patients aged 0 to 4 years. The results showed a different pattern of factor loadings such that we named these factors “provider competence,”“caring and kindness,” and anxiety and pain (Table 3). The three factors were significantly correlated with both procedural and visit satisfaction. Provider competence was the only significant predictor in the logistic regression model predicting excellent satisfaction with the procedure. Provider competence and caring and kindness predicted excellent satisfaction with the visit. After controlling for patient age, parent age, race, ethnicity, and education level, there was no change in the pattern of significance for the factor scores (Table 5).
|Predictors||Adjusted OR (95% CI)*‡|
|Procedure Satisfaction||Visit Satisfaction|
|Provider competence||8.0 (3.7–17.2)†||8.9 (3.4–23.2)†|
|Caring and kindness||1.7 (1.0–3.0)||3.9 (1.7–9.0)|
|Anxiety and pain||1.0 (0.6–1.7)†||1.1 (0.6–2.0)|
White parents were more likely to give excellent ratings of visit satisfaction in both the main analysis (adjusted odds ratio [OR] = 2.3; 95% CI = 1.1 to 4.4) and the 0- to 4-years age subgroup (adjusted OR = 7.0; 95% CI = 2.2 to 22.0) relative to all other races. Highly educated parents were less likely to rate the visit as excellent in the 0- to 4-years age group (adjusted OR = 0.2; 95% CI = 0.04 to 0.9).
To the best of our knowledge, this is the first study to use factor analysis to study satisfaction with emergency care in a pediatric population. There are previously reported studies in adults.11,14 It is also the first to determine elements of parental satisfaction with laceration repair in a pediatric population. Prior studies on satisfaction with laceration repair have focused only on the long-term cosmetic outcome15,16 and have not taken into account the direct ED experience of the family.
Our analysis resulted in an easily interpretable set of factors that make clinical sense and are also significantly correlated with both outcomes. However, the logistic regression showed that not all the factors were significant independent predictors of the outcomes. The provider performance factor was the strongest predictor of excellent visit satisfaction and excellent satisfaction with the procedure itself. This factor included the majority of the domains that have been shown previously to predict patient visit satisfaction: explanation provided to patients,3,4,17,18 caring attitude of nurses and physicians,2,18 and patient perception of provider competence.19 In previous reviews, these domains have been summarized by the term “provider interpersonal skills.” In fact, provider interpersonal skills has been the most replicated predictor of ED patient satisfaction.20,21 Our results add support to these findings regarding the importance of this factor in contributing to satisfaction with care.
We found it interesting that our primary outcomes both had different predictors. The predictors for excellent satisfaction with the procedure were provider performance and cosmetic appearance, while for excellent satisfaction with the visit, provider performance was the only predictor. It is reasonable to expect that parents would rate the procedure not just by provider specific qualities alone but also by some measure of the immediate cosmetic outcome of the wound, since that is, after all, the reason they sought care. But for satisfaction with the visit, cosmetic appearance lost significance as a predictor, leaving just provider performance and hence mirroring previously referenced studies on satisfaction where the primary outcome was overall visit satisfaction. It should be noted that only two variables comprised our cosmetic appearance factor, and the internal consistency for this factor was moderate. Thus, conclusions drawn from analyses of this factor ought to be tempered until replicated.
The subgroup analysis for the infant and toddler group (0 to 4 years) resulted in different factors being generated, albeit with the same theme as the overall factor analysis. The caring and kindness factor was the main difference. This factor comprised two variables (caring attitude of physician and caring attitude of the nurse). This suggests the importance of this specific aspect of the provider–patient interaction with regard to the care of the youngest patients. Parent self-reported ratings of their anxiety and their ratings of the patient’s anxiety were higher compared to ratings in the older patients (p < 0.05), supporting our clinical experience. This observation may actually have contributed to the emergence of caring and kindness as a separate factor in this subgroup.
The anxiety and pain factor was not an independent predictor of either satisfaction with the visit or satisfaction with the procedure itself, a result that may seem surprising at face value, given that laceration repair is a potentially painful and anxiety-producing procedure. It has, however, been documented that satisfaction may relate not to the severity of pain experienced, but rather to the patient’s belief that the medical team has the intention of treating the pain.22 This finding is aligned with the domain of provider communication and caring attitude. Some studies have failed to document any relationship between pain control and patient satisfaction,23,24 while others have found that there are increased satisfaction rates with adequate pain control.18,25 Nonetheless, pain control is a definite patient priority, and of note, during the study period, the majority of patients received topical anesthesia (lidocaine-epinephrine-tetracaine) and many had oral anxiolysis with midazolam.
In our study, perceived length of stay did not load strongly on any of the factors and so was not a part of the final factor analysis. The median length of stay for patients treated for lacerations during the study period was 173 minutes, which is comparable to national length of stay data for pediatric EDs.26 While several studies have found length of stay to be an important determinant of satisfaction,5,6 in our study the parent’s perception of this variable did not load strongly enough on any of our factors to warrant inclusion. Thus while length of stay is clearly important to the satisfaction of parents and patients, it may be independent of parents’ perceptions of other aspects of their medical care experience during a visit for laceration repair.
This was a single-center study, which may limit generalizability, since our institution may differ from others with respect to laceration care practices and parent demographics, especially with regard to the high education level of the majority of respondents. Being a convenience sample, it may not have been representative of all the patients who presented during the study period, and although we had a comparatively high response rate of 76%, the nonrespondents may have differed sociodemographically or in their opinions of the care received, especially the most dissatisfied parents who may have been less likely to spend time completing a survey.
We used a very high standard for measuring satisfaction using a cut-point of excellent versus other. With previously known high satisfaction scores in our ED, differentiating excellent care from other ratings was important for determining provision of the highest quality patient-centered care. In addition, we classified any survey with missing questions as incomplete, and we excluded them from the factor analysis (15% incomplete). Most were incomplete because parents inadvertently skipped pages. Although this may have influenced the factor analysis, we decided a priori to adopt this conservative approach to manage missing data.
We did not measure the parents’ rating of discharge instructions because enrollment occurred while awaiting formal discharge and instructions. This could have influenced the factor analysis, but we felt that waiting until after discharge instructions were provided would have had a detrimental effect on parent enrollment. Completing the survey at the point of care could have biased the responses in a favorable direction, although we emphasized to parents the confidentiality of their responses. Finally, we did not resurvey for long-term cosmetic outcomes because we were interested in immediate satisfaction with the ED visit.
Our study shows that in a population of parents presenting with their children specifically for laceration repair, high parent ratings of provider-specific qualities as summarized by our provider performance factor are the strongest predictor of excellent parent satisfaction. Our finding suggests that from the parents’ perspective, providers who pay close attention to provider–patient interactions by trying to communicate effectively, portraying a caring attitude to the family, and displaying professionalism in their actions positively affects satisfaction scores and may result in more highly satisfied parents.
The authors acknowledge Elizabeth Paulsen and Lucy Abernethy for their invaluable assistance in data collection and data entry.