Instruments to Measure Perceptions in the Emergency Department Provider- Patient with Sickle Cell Disease Interaction: Findings of an Integrative Review from a Ph.D. Project

Background & Purpose: The purpose of this review is to examine instruments that measure providers' perceptions of adult patients with Sickle Cell Disease (SCD), examine instruments that measure adult patients with SCD perceptions of providers' behaviors, and determine optimal instruments to use in evaluating the perceptions of Emergency Department (ED) providers and adult patients with SCD of one another's behaviors after an interaction in


INTRODUCTION
Sickle Cell Disease (SCD) is the most common inherited blood disorder in the United States [1]. During childhood and adolescence, many patients with SCD have disease-related symptoms treated at pediatric comprehensive sickle cell clinics [2]. Once patients with SCD reach adulthood, they transit out of comprehensive sickle cell clinics into mainstream medical pain, increased risk of infections, and eventually end-organ damage [13 -15]. Patients with SCD can also experience intermittent acute pain from an acute Vaso-Occlusive Crisis (VOC), which is the hallmark symptom of SCD, leading to an ED visit [10,16,17]. Pain from acute VOC accounts for 78% of ED visits made by adult patients with SCD [18].
Treatment for an acute VOC includes evidence-based pain management with a combination of nonsteroidal antiinflammatory medications and short-and long-acting opioids used in-home therapies [19]. When home regimens fail, aggressive analgesic management, using parenteral [intravenous or subcutaneous] opioids, should be initiated within 30 minutes of triage or 60 minutes from arrival to the ED [19,20]. Patients with SCD wait for 25% longer to see a physician than other patients presenting to the ED [21] or 16 minutes longer when the average wait time is one hour [22]. The average wait time to initial analgesic administration is 80 minutes from the time of triage [23]. This long wait time for care has been associated with race [21]; over 90% of individuals with SCD in the United States are African American [24]. African Americans with chronic diseases suffer more negative health outcomes leading to substantial health disparities in timely access to quality care [25].
A recent survey of ED Providers reported that 75% of providers were not aware of evidence-based guidelines regarding patients with SCD, although 87% were aware that pain medication should be administered within 30 minutes of arrival to the ED [26]. Providers have reported reluctance in following guidelines due to the potential for drug addiction or suspected drug-seeking behaviors among patients with SCD when patient behaviors such as watching television or talking on the telephone seem incongruent with subjective pain scores [27 -29]. It is common for patients with SCD to present with no outward signs of pain [30]. Physicians may expect to see signs such as redness and swelling of a joint related to the acute inflammatory process often associated with a VOC [30]. When providers do not find outward signs of pain in a patient with SCD, they report feeling frustrated and manipulated [30,31]. These feelings of frustration and manipulation may increase, particularly when the provider reviews the medical chart and finds out that a patient has had many visits to the ED for pain and is non-adherent with medical care, such as following up with a PCP after discharge from the ED [30].
Patients with SCD are often reluctant to go to the ED for analgesic pain management due to anticipated stigma based on past negative experiences with providers [28,32]. Patients with SCD often report feeling health-related stigma, or feelings of being discredited based on their health condition when providers doubt their quest for pain relief and label them as being drug-addicted, drug-seeking, or manipulative [33,34]. Patients with SCD also perceive stigma when they are treated differently than other patients, making them feel undervalued and not a part of their health care decisions [34,35]. The quality of communication with a provider can be directly associated with the ability of the patient with SCD to trust a provider during the interpersonal interaction [36].
Many providers are unaware that some patients with SCD perceive responses to their behaviors as unfavorable and stigmatizing during an interaction [28, 37 -41]. Patients may not understand that behavioral attributions, such as requesting specific dosages of opioid pain medications or frequent visits to the ED for pain management, can lead to cognitive bias, negative behaviors and stigmatization by providers [17,42,43]. This discordance in the patient-provider interaction causes a communication breakdown that affects patient outcomes and may have significant clinical implications [39].
Patient-provider concordance, or a mutual understanding, in a medical interaction, can improve health outcomes [8, 44 -48]. There is no gold standard to measure concordance (or discordance) in the shared experience of the provider and patient interaction [44,49]. While it is challenging to measure perceptions of patients' and providers' self-reported agreement of an interaction, a paired survey method has previously been successful in measuring the patient-provider interaction for concordance (or discordance) and examining the extent of shared perceptions of behaviors during an interaction [44, 46, 49 -52]. A dyadic or paired survey approach has not been used in current studies measuring perceptions in the provider and patient with SCD interaction. Current studies utilize instruments that measure provider behaviors such as negative attitudes and the patients' perception of health-related stigma during separate encounters [37, 53 -66].
An examination of provider and patient with SCD perceptions of behaviors during the same ED encounter is needed in order to bring awareness to potential discordance in the patient-provider interaction and develop strategies to reduce possible discordance [57, 67 -69]. A valid and reliable way to assess provider and patient with SCD perceptions of behaviors during the same ED encounter needs to be established in order to evaluate this interaction for possible discordance. Therefore, the purpose of this review is to: (a) identify and critically appraise instruments used to measure providers' perceptions of adult patients with SCD behaviors, (b) identify and critically appraise instruments used to measure adult patients with SCD perceptions of providers' behaviors, (c) describe the psychometric properties of these instruments, and (d) determine optimal instruments to use in a paired-survey method to evaluate the dyad immediately after an interaction in the ED.

Search Strategy
EBSCOhost and PubMed databases were searched in June 2020 after consultation with a medical reference librarian. EBSCOhost includes Cumulative Index to Nursing and Allied Health Literature (CINAHL), MEDLINE, and PsychInfo databases. All databases were searched for articles reporting the psychometric properties of instruments using the following synonyms of key terms: instrument [OR] (Fig. 1), which follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines, provides a flow chart of the search strategy selection process that was used to identify relevant articles for review [70].
The search yielded 146 articles from EBSCOhost and 90 from PubMed. After 90 duplicates were removed, title and abstract review were conducted on 146 articles. Articles were excluded if they did not include patients with SCD aged 18 or older (n=98), did not include a licensed health-care provider (i.e., Medical Doctor, Doctor of Osteopathic Medicine, Registered Nurse, Physicians Assistant, or Nurse Practitioner) and patient with SCD interaction for pain (n = 36), were not conducted in the United States (n=1), or were non-English (n = 1). Title and abstract review yielded ten articles. A secondary search was conducted within EBSCOhost, PubMed, Google Scholar, and in a hand search of references using an ancestry and descendancy method of instruments identified in the ten articles utilized previously on a different population (i.e., child or adolescent with SCD, internal medicine resident, inpatient nurse), or in a different setting (i.e., inpatient, provider's office). Using ancestry and descendancy methods to identify instruments helped to discover the steps used to develop an instrument and subsequent uses of an instrument to assist in the psychometric property critique of an instrument. Eleven additional articles were included using this ancestry and descendancy method to identify papers describing an instrument's development or subsequent use.  Trust & discrimination (n = 1) Problematic hospital experiences (n = 1) Optimism & perceived discrimination (n = 1) Perceived discrimination & burden of pain (n = 1) Provider reaction to medical conditions (n = 1) During the full-text review, 21 articles were examined to identify instruments that specifically measured a provider's perceptions of a patient with SCD behaviors or a patient with SCD perceptions of provider behaviors. The specificity of the instrument used in a population was chosen because instruments have been shown to have lower validity when administered outside the context for which they were developed [71]. For the purpose of this review, the behaviors associated with stigma, stereotype, prejudice, bias, attitude, discrimination, and racism, were targeted due to the high use of these terms in SCD behavioral research. While the root causes of behaviors associated with these terms are different, they are psychosocial stressors and follow similar social processes in a person's perception of them [72]. Reasons for exclusion during the full-text review included instruments not explicitly designed for the provider or patient with SCD interaction but were designed to measure patient trust in communication with a provider [36], patient trust and discrimination [73], patient experiences in the hospital as being problematic or not [74], patient optimism and perceived discrimination [75], the association of perceived discrimination with the burden of pain [76], and an instrument designed to measure provider reaction to medical conditions [65]. No date limit was set for this literature search. After full-text review, the resulting 15 articles reporting on six instruments were included in this integrative review. An overview of the 15 articles included in this review is summarized in Table 1. There is no gold standard to evaluate and interpret the quality of studies included in an integrative review, which includes various research designs [78]. Therefore, the studies presented in this paper were reviewed if they met the inclusion criteria, and no studies were excluded due to methodological criteria [79]. Heterogeneity was not assessed as the difference between instruments is under investigation, not the methodology of the studies under which they were published. To prevent publication bias in this review, multiple databases were searched, web searches conducted, ancestry and descendancy handsearching methods utilized, and grey literature included [80]. This review did not include statistical evaluation of individual studies, therefore, additional methods to prevent publication bias, such as funnel plots, were not explored [78].

Psychometric Properties
The psychometric evidence strength for each instrument included in this review was evaluated using the Psychometric Grading Framework (PGF), which was developed to evaluate self-report instruments by determining quality across six psychometric properties: content validity, construct validity, criterion validity, internal consistency, test-retest reliability, and inter-rater reliability [81]. The feasibility of each instrument was also assessed to determine the instruments' ease of use and potential administration limitations.
The PGF consists of two scales. Scale 1 is a matrix from which each psychometric property of an instrument is assigned a grade, A-D, with A at the top of the hierarchy and D at the bottom of the hierarchy. A higher grade reflects a more reliable psychometric property [81]. The graded hierarchy strength of each psychometric property was generated from the recommendations of commonly used guidelines for threshold values of statistical tests [81]. The determination of the graded hierarchy strength of each psychometric property within an instrument can be visualized in the Level of Psychometric Measures (  Any variable which appropriately matched a psychometric property was included in the grading of that psychometric property. Grades are not additive or averaged within a psychometric property and only assist the researcher in determining an overall grade for the psychometric property. For the purposes of this review, if multiple grades were reported for a psychometric property, the grades were assigned a number for a letter grade (A = 4, B=3, C=3, D=1, F=0) and the result averaged and rounded for determination of that psychometric property's grade. Scale 2 ( Table 3), Grading of Psychometric Strength, reviews the overall strength of the instrument is determined with the use of and the subsequent assignment of value: "good," "adequate," "weak," or "very weak" [81]. The process of determining the overall strength of an instrument can be visualized in the Grading of Psychometric Strength (Table 3) [81]. It is up to the individual researcher to determine the PGF grade of psychometric strength level acceptable for instrument use in research [81].

RESULTS
The results of this review identify instrument characteristics, instrument purpose and development, psycho-metric properties, and feasibility of each provider-and patient-focused instrument. Table 1 provides a summary of the 15 studies included in this review.

Instrument Characteristics
This review included 15 articles representing six instruments. Articles were published between 2009 and 2018, and all were conducted in the United States. Twelve articles reported provider-focused instruments (n = 4) and three articles patient-focused instruments (n = 2). Table 4, Provider-focused instrument overview 1, provides an overview of providerfocused instrument name purpose, theoretical framework, sample demographics, and setting., and Table 5, Providerfocused instrument overview 2, provides the study design or sampling technique, number of items per subscale, scoring method, and any additional instruments, variables evaluated, or interventions discussed in provider-focused studies. Table 6, Patient-focused instrument overview 1, provides an overview of patient-focused instrument name purpose, theoretical framework, sample demographics, and setting, and Table 7, Patient-focused instrument overview 2, provides the study design or sampling technique, number of items per subscale, scoring method, and any additional instruments, variables evaluated, or interventions discussed in patient-focused studies.

Good
Three or more As and/or Bs ± C or D A + A + B + C

Adequate
Two As and/or Bs ± C or D A + B + C + D

Weak
One A or B ± C or D B + D + C

Very weak
One or more C or D only D + C  Pre-experimental one group pre-test/post-test quality improvement project 10 items. 5-point Likert scale: 5 -much more than average to 1 -much less than average, 1strongly agree to 5 -strongly disagree, or 5 -extremely likely to 1 -not at all likely.   Provider-focused instruments were reported to measure providers' attitudes [35, 37, 54, 55, 57 -61, 64, 66, 77]. Provider behaviors of interest were most often adapted from extant literature and reviewed by an expert panel of providers developed by authors of a manuscript [37,58,66]. Expert panels most often included a provider experienced in treating patients with SCD, providers experienced in the measurement of the patient experience of care, providers experienced in bioethics, and a patient with SCD to gain the patient perspective [37,58,66]. Behaviors of interest often included a provider's regard for a patient, feelings of affiliation with a patient, the level of frustration a provider had with a patient, beliefs providers had about a patient's behavior, and behaviors providers associated with less than favorable attitudes toward a patient [58,66].
Patient perceptions of provider behaviors were evaluated by measuring health-related stigma [53,62,63]. Health-related stigma occurs when a health-related characteristic of a patient evokes a negative reactive behavior from a provider such as a negative emotion, negative attitude, rejection, devaluation, blame, avoidance, or exclusion [33] from a person in a position of authority [83].
All of the instruments were administered through selfreport using 4 to 10-point Likert-style scoring systems for between 7 to 31 items per instrument. The largest group to complete an instrument were located at a conference (n=795] [35]. All patient-focused instruments were administered in a SCD clinic or SCD center [53,62]. No instruments were administered to both patients and providers to evaluate their perceptions of an interaction during or immediately after the encounter.
While a theoretical framework can standardize and guide the development and measurement of related concepts or variables, only four articles reported the association of a theory during the administration of an instrument, one of which was a secondary analysis [54,60,62,63]. The Theory of Self-Care Management for Sickle Cell Disease was used in the administration of a patient-focused instrument in two articles and a provider-focused instrument in another article [60,62,63]. One theory was used in the administration of a providerfocused instrument, Lewin's Change Management Model [54]. No theoretical frameworks were applied to the development of instruments.

Provider-Focused Instruments
All provider-focused instruments were designed to measure provider attitudes towards patients with SCD. Provider-focused instruments measured a provider's perceptions of patients' with SCD behaviors through the measurement of positive attitudes toward patients, negative attitudes toward patients, perception of patients' behaviors that raise a "red flag" for the provider, perceptions of patients' behaviors that were concerning to the provider, and in some cases the level to which observation of another providers behaviors toward patients' with SCD bothered the provider [35, 37, 54, 55, 57 -61, 64, 66, 77]. No provider-focused instruments measured stereotype, prejudice, discrimination, or racism directly. Jenerette et al. (2015) adapted the General Perceptions about Sickle Cell Scale to include three stigmarelated items from a patient-focused instrument to assist in measuring nurse attitudes [60]. The PASS instrument was used to associate physician attitudes with "stigmatizing language" by evaluating a provider's charting language [55]. This charting language consisted of terms that can create doubt in the reader, such as a provider charting the patient's pain score was "still a 10" instead of "still has 10/10 pain" or using the term "narcotic" instead of "opioid" [55]. Freiermuth et al. (2014) included an additional instrument to evaluate provider biases, emotions, and expectations, but the additional instrument was not developed to evaluate providers who care for patients with SCD [57]. Ratanawongsa, et al. (2009)  The 10-item PASS questionnaire was initially developed using the hypothesis that providers may have believed some patient behaviors, such as the exaggeration of pain, manipulation, drug abuse, and non-compliance, were associated with provider attitudes toward patients with SCD [66]. Provider perceptions of a patient with SCD behavioral attributions may influence a provider's interpretation of patients' behaviors in a negative way [66]. The PASS was used to measure ED provider attitudes after education and the introduction of a pain algorithm specific to patients with SCD [54]. PASS was later revised with the addition of 3 questions of the 10-item PASS, to measure the influence of language used in provider charting was compared to provider-in-training attitudes [55].
Content was added to PASS by  when developing the Clinician Attitude Scales. This content included questions to measure provider positive attitudes, provider negative attitudes, provider perceptions of concernraising behaviors exhibited by patients, and provider perceptions of red-flag behaviors exhibited by patients [58]. In a secondary analysis of content within the Clinician Attitude Scales, the 31-item instrument was modified into a final 17item questionnaire to measure negative attitudes, positive attitudes, and concern-raising behaviors [57].
The Clinician Attitude Scales were adapted by an expert panel consisting of ED providers who treat patients with SCD, a hematologist specializing in SCD, and a patient with SCD [35]. This new General Perceptions About Sickle Cell Patients Scale was administered by  to ED providers to measure negative attitudes, positive attitudes, and red-flag behaviors in order to correlate instrument results with provider practice patterns [35]. The results from the newly adapted scale were also correlated with the use of the derogatory term 'sickler,' which is a term sometimes used by providers as a label for patients with SCD [37]. This version of the General Perceptions About Sickle Cell Patients Scale was also used to measure nurses' attitudes at two different worksites [60], to determine if educating providers about SCD would change provider knowledge and attitudes toward patients with SCD [61], and improve ED providers' attitudes toward patients with SCD in pain [59].
While all other provider-focused instruments were administered to samples, which included ED providers, only one was adapted for the ED provider in the ED environment [57].  revised the initial 17-item Clinician Attitude Scales were by adding four items on provider frustration and provider perceptions of addiction among patients with SCD [57]. One item was removed due to a lack of relevance to the ED environment. Items were assessed by content experts to determine which items would be relevant for ED providers in the ED environment [57]. After factor analysis, a new 18-item General Perceptions About Sickle Cell Disease Patients Scale was developed to measure ED provider behaviors [57]. Subscales included negative attitudes, positive attitudes, and uneasiness with care in the ED environment [57].

Patient-Focused Instruments
Patient-focused instruments measured perceptions of providers' behaviors through instruments to measure healthrelated stigma [53,62]. No patient-reported instrument measured stereotypes, prejudice, biases, attitudes, or racism directly. One instrument used the term 'expected discrimination' with a definition including 'anticipated discrimination' alluding to anticipated stigma in addition to measuring feelings of 'social exclusion' and 'disclosure concerns' [53].
The two patient-focused instruments include a 30-item Sickle Cell Disease Health-Related Stigma Scale (SCD-HRSS) [62,63] and 11-item Measure of Sickle Cell Stigma Scale (MoSCS] [53]. Both patient-focused instruments were developed to measure how patients with SCD perceived healthrelated stigma [62]. Health-related stigma instruments measured how patients with SCD perceived stigma from doctors, the public, and family members [62]. Jenerette, et al. (2012) developed the SCD-HRSS to measure external stigma, which is an individual's awareness of unfair treatment or devaluing attitudes towards them [53,84]. The SCD-HRSS also focuses on perceived stigma and experienced stigma through items wherein a patient is asked to reflect on incidents after they have occurred [53,62]. To address a gap in measuring internalized stigma and anticipated stigma of a patient with SCD, the MoSCS was developed by Bediako, et al. (2016). The MoSCS was developed by adapting a 40-item HIV stigma scale [53]. Focus groups of patients with SCD were asked to score the items as very relevant or not at all relevant to the experiences of patients with SCD [53]. The resulting 31-item MoSCS was administered to a convenience sample and after analysis, and 11 items were retained in the scale. The MoSCS applies a multi-dimensional approach to capture a patient's social exclusion, expected discrimination, disclosure concerns, and internalized stigma based on previous experiences and possible future experiences [53]. The MoSCS is designed to capture the lived experience of a patient and does not focus on specific health interactions [53].

Psychometric Properties
An overview representation of the psychometric properties for instruments based on analysis using PGF criteria can be found in Table 8, level of psychometric measures for providerfocused instruments; Table 9, grading of psychometric strength for provider-focused instruments; Table 10, feasibility found in provider-focused instruments; Table 11, level of psychometric measures for patient-focused instruments; Table 12, grading of psychometric strength for patient-focused instruments; Table  13, feasibility found in patient-focused instruments. Of note, no studies reported test-retest no inter-rate reliability for the provider-or patient-focused instruments, so are not included in the tables for the level of psychometric measures. In addition, currently, no gold standard instruments are available, which could be tested against instruments for criterion validity, so they are not included in the tables for the level of psychometric measures. Higher patient socioeconomic status (p <0.001) would equal more positive provider attitudes. 3. The more severe the SCD was using as evidence by increasing the frequency of hospitalizations within the past year (p<0.001), the less positive attitudes. 4. concerning behaviors such as patients discharge against medical advice (p = .04), narcotics dependence (p = 0.01), and positive toxicology screen (0.04) would equal less positive provider attitudes. 5. Prior disputes with staff (p=0.002) would equal less positive provider attitudes. 6. No statistical significance was found in the sex of the provider. Two markers of SCD did not meet statistical significance (prior avascular necrosis and prior acute chest syndrome). Exploratory Factor Analysis: Eigenvalues above 1 = 2, retained due to lower AIC for 2-factor solution. PGF Grade = A. Corrected item total correlation ranges 10-item/7-item = 0.54 -0.80/0.59-0.81. 3 items eliminated due to high uniqueness values on factor analysis. PGF Grade = A.

PASS PGF Grade = C.
Hypothesis: 1. PASS scores differ in stigmatizing language vs. neutral language (p<0.001) in medical students and residents. 2. Stigmatizing language cohort prescribed less aggressively (p<0.001) and had less comfort in dosing pain medication (p<0.04) than neutral cohort. 3. The more training a physician had, the lower the PASS score (p<0.001). Perceptions of Vignette Physician attitudes: stigmatizing language note physician had a more negative attitude toward the patient (p<0.001).
Hypothesis: Physicians who use the term 'sickler' have more negative attitudes towards patients with SCD (statistically significant, p-value not given) and less likely to adhere to national guidelines (not statistically significant). PGF Grade = A.

General Perceptions About Sickle Cell Patients Scale [60]
Reported Clinician Attitude Scales. PGF Grade = C.
Reported previous construct validity with correlation to MCRS. PGF Grade = A. Reported previous construct validity with SCD-HRSS correlation to BDI-FS. PGF Grade = A. Hypothesis: ED/ICU nurses and medical-surgical nurses demonstrate different attitudes and behaviors toward patients with SCD. ED/ICU and more negative attitudes (p = 0.342), concern raising behaviors (p=0.232), red flag behaviors (p =-0.186), and lower positive attitudes (p = 0.661). PGF Grade = D. Hypothesis: Attitudes and behaviors vs. appropriate use of pain medication (negative p < 0.001, positive p = 0.015, concern-raising behaviors p < 0.0001, red-flag behaviors p = 0.002), comparing patients with SCD to other patients with medical conditions (negative p < 0.001), concern-raising behaviors p = 0.005, red-flag behaviors p = 0.012. PGF Grade = A.

Overall PGF Grade = A + A + A + D + A = B.
Reported previous Cronbach's alpha scores ranging from 0.76 to 0.89 for negative attitudes, positive attitudes, concern-raising behaviors, and red flag behaviors. PGF Grade = B.
Reported total score and subscales of SCD-HRS as ranging from 0.69 to 0.84. PGF Grade = B. Overall PGF Grade = B+B = B.

General Perceptions About Sickle Cell Patients Scale [61]
Reported Clinician Attitude Scales. PGF Grade = C.
Hypothesis: Change in clinician SCD knowledge and attitudes toward patients with SCD before the intervention, immediately post-intervention, and 2 months post-intervention. Overall increase in knowledge scores (p<0.0001), increase in overall knowledge between T1-T2 (p<0.001), increase in overall knowledge between T1-T3 (p = 0.0125) Total negative attitudes decreased between T1-T3 (p = 0.03). PGF Grade = A.
Reported previous Cronbach's alpha scores ranging from 0.76 to 0.89 for negative attitudes, positive attitudes, concern-raising behaviors, and red flag behaviors. PGF Grade = B.

General Perceptions About Sickle Cell Patients Scale [59]
Reported Clinician Attitude Scales. PGF Grade = C.
Reported Cronbach's alpha from previous administration Hypothesis: 1. When compared to physicians, nurses had higher levels of frustration in caring for an adult patient with SCD (p = 0.0037), higher perception of opioid addiction in the general population (p ≤ 0.0001), higher perception of opioid addiction among adults with SCD (p ≤ 0.0001), and a higher perception of ED patients with SCD believed to be addicted to opioids (p = 0.0210. In all cases, ratings of frustration decreased as the years of practice increased. PGF Grade = A. Partial Spearman correlation with subscales and MCRS total scores.  Reported Clinician Attitude Scales. PGF Grade = C. Additional questions validated by content experts. PGF Grade = C.

Instruments to Measure Perceptions in the Emergency
Overall PGF Grade = C+C = C.
Hypothesis: ED provider attitudes would improve after an intervention and be maintained over a 2.5 year time period at two sites by attitude subscale scores, level of frustration, and perception of addiction among adults with SCD compared with the general population and ED patients. Uneasiness with care increasing as the years of provider experience increased (p = 0.0221); negative attitudes decreased between time 1 and time 2 (p = <0.0001); nurses at site 2 higher negative attitudes vs. site 2 physicians, site 1 nurses and physicians (p < 0.0001); site 2 physicians higher negative attitudes vs. site 1 nurses and physicians (p<0.04); average negative attitudes higher at site 2 than site 1 (p<0.0001); uneasiness with care higher in physicians than nurses (p<0.0001); positive attitudes increased over time (p=0.0324); positive attitudes increases between time 1 and time 3 (p=0.0134); positive attitudes higher at site 1 than site 2 (p<0.0001); frustration level higher at site 2 than site 1 (p<0.0001); frustration level higher in nurses than physicians (p = 0.0015); site 2 estimated higher addiction than site 1 (p<0.0001); nurses estimated higher addiction than physicians (p<0.0001); nurses at site 2 estimated higher addiction than nurses and physicians at site 1, and physicians at site 2 (p<0.0001); physicians at site 1 estimated lower addiction than physicians at site 2 (p=0.0174 and nurses at site 2 (p<0.0001). PGF Grade = A.    Adapted from a previously validated Chronic Pain Stigma Scale by changing the words chronic pain to sickle cell pain or sickle cell disease, process not listed but assumed. PGF Grade = D.
Measured against BDI-FS = significant but weak correlation (r = 0.27, p < 0.05). Did not conduct factor analysis due to small sample size. PGF Grade = C. Reports adaptation from a previously validated Chronic Pain Stigma Scale by changing the words chronic pain to sickle cell pain or sickle cell disease, process not listed but assumed. PGF Grade = D.
Hypothesis: Patients w/ SCD who receive the intervention will have lower perceived healthrelated stigma from doctors compared to those in the control group measured by mean change from baseline (Time 1) on health-related stigma total score (p = 0.2297) and doctor subscale score (p = 0.0002) by Time 3. PGF Grade = A.   After evaluation with the PGF, all provider-and patientfocused instruments had 'adequate' psychometric strength, the second-best overall strength in this appraisal tool [81]. One article reported adaptation of items within the PASS instrument, which led to that administration of the instrument being rated as 'weak' due to no internal consistency measurements provided for the altered questions [55]. The most consistent psychometric property reported was Cron-bach's alpha as a measure of internal consistency. All but one article reported hypothesis testing. Content validity was determined by authors of manuscripts through literature reviews, provider expert panels, provider expert panels with patient perspectives, and a content validation survey. All but one instrument was rated a 'C' for content validity according to the PGF. The SCD-HRSS was the only scale rated a 'D' in grading strength for content validity, according to the PGF. Construct validity was demonstrated through hypothesis testing, factor analysis, or inter-item and total-item correlations. All instruments included in this review received an 'A' grading strength for construct validity based on the PGF. The PASS 10-item scale reported the highest overall internal consistency (Cronbach's alpha = 0.91] [66]. Subsequent iterations of the use of PASS items reported subscale internal consistencies only, with the provider negative attitudes subscale reporting the highest internal consistency of subscales (Cronbach's alpha = 0.91-0.93). The 11-item MoSCS reported the highest overall and subscale internal consistency (Cronbach's alpha = 0.87 overall; 0.89-0.76 subscales] and was found to be the most multi-dimensional instrument incorporating multiple types of patient-perceived stigma [53]. Criterion validity could not be measured due to a lack of a gold standard comparison for the concepts or constructs measured in this population. The only instrument administered to ED providers in the ED environment was The General Perceptions About Sickle Cell Disease Scale [57]. This instrument was considered 'adequate' according to the PGF. Freiermuth et al. (2014) based content validity on the previously reported Clinician Attitude Scales and had additional questions validated by content experts earning the highest grade of 'C' for this psychometric property. The General Perceptions About Sickle Cell Disease Scale and subscales were tested using a partial Spearman correlation with MCRS total scores and had a construct validity grade of 'A,' the highest grade for this psychometric property. The sample size was at the low end of acceptable for the use of the instrument to associate provider characteristics with provider attitudes (n = 215) and to evaluate the effectiveness of an intervention (n=216), potentially affecting the stability of psychometric property results. Internal consistency was quite high for the negative attitudes subscale (Cronbach's alpha = 0.93) and acceptable for uneasiness with care (Cronbach's alpha = 0.83) and positive attitudes (Cronbach's alpha = 0.82) subscales, earning an overall grade of 'A.' The authors found an additional dimension not identified in the original administration of the instrument, administered to non-ED providers, which warranted the adaptation of the scale to meet the needs of evaluating ED providers.
Only the Doctors subscale of the SCD-HRSS met criteria for potential use in the ED environment with ED providers as the Public and Family subscales are not the populations of interest in this review [62]. The SCD-HRSS was measured against the BDI-FS for construct validity, which showed a significant but weak correlation; this suggests that those with higher levels of perceived health-related stigma have higher levels of depression. The internal consistency was acceptable with subscale Cronbach's alpha ranging from 0.68 to 0.82 and total scale Cronbach's alpha of 0.84. The Doctors subscale (Cronbach's alpha = 0.68) was hypothesized by the authors to be low due to the wording of doctor-related questions. Another factor that may affect the psychometric properties of this instrument is the small sample size for instrument development (n = 77) and for instrument use in an intervention (n = 90). A more recent iteration of the SCD-HRSS to develop a 'nurses' subscale was administered to youth and demonstrated an overall consistent face validity of 90.5% [82 -85], but was not included in this review due to the age of the population being evaluated.

Feasibility
Ease of use of an instrument was noted through administration time, the method of instrument administration, and the availability of an instrument. Four articles were secondary analyses and thus were not included in the feasibility assessment [37,59,63,77].
Only Jenerette et al. (2012) reported a relatively short patient-focused instrument administration time of 30 minutes [62]. No administration time estimates were reported for provider-focused instruments [35, 37, 54, 55, 57 -61, 64, 66, 77]. The method of administration was more likely to be written [35,54,57,58,62,64,66] than electronic [55,60,61], although two studies started with paper and pencil and moved to electronic administration during follow-up to increase confidentiality and reduce the effects of social desirability bias [53,64]. All have the capability for electronic administration.
Instruments were readily available in the appendices of four articles [35,54,58,59], and instrument questions from three other articles were available in tables reporting psychometric properties but would require formatting prior to administration [53,64,66]. Two articles reported some questions in the literature [55,60], and two articles did not provide questions, although the authors refer the reader to the original developer of the instrument [61,64]. No articles reported the need to request permission to use an instrument or a cost to use an instrument. After contact with Jenerette et al. (2012) [62], the SCD-HRSS instrument questions and scoring were available after completion of a permission to use form.

CONCLUSION
The purpose of this integrative review was to identify and critically appraise instruments used to measure perceptions of behaviors in the provider and patient with SCD interaction, describe the psychometric properties of these instruments, and recommend optimal instruments to use in evaluating ED provider and adult patient with SCD perceptions of each other's behavior during the same interaction in the ED. By identifying and bringing awareness to potential discordant communication and perceptions in the interaction, barriers can be crossed, and health outcomes can be improved through patient-provider concordance in a medical interaction [44,53].
When providers' perceptions of an encounter are measured using an instrument, researchers have found a discordance of provider perceptions when compared to the perceptions of the patient who has just been in the same interaction [86]. A study on dyadic interactions reported the patients' perceptions in discordant interactions were more likely to be that their provider had not listened to them, had left questions unanswered, and did not feel they had input in treatment plans [44]. Providers were also found to underestimate pain during a dyadic interaction [44], which is the primary presenting symptom for a patient with SCD in the ED. Although there is no gold standard to measure concordance [or discordance] in the shared experience [44] of the provider and patient with SCD interaction, a comparative study of both sides of the interaction is needed to understand the extent of shared perceptions during the interaction. While there are several instruments that measure provider behaviors such as negative attitudes and the patients' perception of health-related stigma during separate encounters [35, 37, 53 -65], none have used a dyadic or paired survey approach. In addition, by understanding the shared perspective, interventions can be developed to resolve perceptions of potentially discordant care.
This review found all but one administration of provider-focused instruments had adequate reliability and validity to measure providers' perceptions of patients with SCD behaviors. All provider-focused instruments measured provider behaviors towards patients with SCD. These behaviors included positive attitudes toward patients, negative attitudes toward patients, perception of patients' behaviors which raise a red flag for the provider, perceptions of patients' behaviors which are concerning to the provider, and in some cases the level to which other providers' behaviors toward patients with SCD bother the participating provider. One provider-focused instrument was found to have adequate overall psychometric properties and was conducted in the ED environment with ED providers, the general perceptions about sickle cell disease. This scale was adapted specifically for the ED provider to measure negative attitudes, positive attitudes, and uneasiness with care for the patient with SCD. No barriers in the feasibility of administering the provider-focused instruments were identified.
All patient-focused instruments were found to have adequate reliability and validity. Unfortunately, no patientfocused instrument was administered in the ED environment measuring patient perceptions of ED provider behaviors. One instrument, SCD-HRSS, could potentially be adapted as a comparison instrument due to the inclusion of a Doctor's subscale. A barrier to the use of the SCD-HRSS could be the ability to obtain the questions and the copyright limitations in evaluating the items of the Doctor's subscale.
Directions for future research include comparing perceived behaviors of providers and patients who are part of the same interaction in order to identify concordance (or discordance) in the interaction. No studies to date were identified that evaluated both sides of the provider and patient with SCD interaction during the same encounter. In addition, the inclusion of a theoretical framework can guide and standardize research, and has been used to support the development of concordant care through the process of establishing patient goals, understanding treatment limitations, evaluating treatments received, and judging whether the care received was concordant with a patient's goals, taking into account limitations on treatment [87].
There are multiple limitations to this review. The number of articles identified for review may have been limited by the search strategy, and some instruments may not have been identified. However, several databases were searched on multiple occasions. The keyword search may have been too limiting, even after consultation with a medical reference librarian. The exclusion criteria may have eliminated articles that reported instruments that should have been included, or inclusion criteria could have been broadened to include additional parameters related to the assessment of behaviors. Hand-searching and ancestry/descendancy search methods identified many additional articles included in this review. One author reviewed articles several times to gather all relevant data to include in this review.

RELEVANCE TO NURSING PRACTICE
Communication is successful where there is a mutual agreement or concordance of perceptions during an interaction [44,86,88]. For patients with SCD and patients with other conditions, the ability for the patient-provider dyad to successfully communicate in the ED is vital to achieving positive physical and psychological health outcomes leading to a shared understanding of medical goals, mutual trust, patient satisfaction, and patient adherence [44, 86, 88 -90].
To develop interventions that may improve perceptions within the dyad, the interaction must be investigated using a provider-focused and patient-focused instrument in a pairedsurvey method. A paired-survey method will capture both sides of the interaction. While it is challenging to measure patients' and providers' self-reported perceptions of behaviors, the general perceptions about sickle cell disease patients scale and doctors subscale of the SCD-HRSS would be a valid and reliable way to measure perceptions of behaviors between the patient with SCD and their provider in the ED.

FUNDING
Not applicable.

CONFLICTS OF INTEREST
The authors declare no conflict of interest, financial or otherwise.