Development and initial validation of the addressing client needs with social determinants of health scale (ACN: SDH)

SDOH competence and understanding the relationship to health and wellbeing is critical for a practitioner to address SDOH needs in practice [6, 7]. Research has shown that physicians continue to lack sufficient training in SDOH, which minimizes their ability to address SDOH needs in practice [6]. In a qualitative study of pediatric nurse practitioners, observers found that the trainees were able to identify some social health needs but very few intervened [8]. Feedback from the two cohorts of pediatric nurse practitioners (N = 36) found that trainees expressed discomfort in asking SDOH questions that they deemed sensitive and they experienced difficulty in negotiating their priority to address the medical need versus the non-medical social health need [8]. In a study of civil servants (N = 153), researchers found that only 50% of the respondents were familiar with SDOH, 83% indicated they need more information on effective intervention strategies [9]. Similarly, another study in Canada investigated the relationship between understanding and attitudes toward SDOH with people who worked at community-based organizations [10]. Results found that support for addressing health inequalities was associated with increased awareness of SDOH and a broad view of the influence of SDOH on overall health and wellness [10].

Lastly, healthcare systems are eager to effectively address SDOH needs of patients because of the positive impact on population health, healthcare costs, and overall patient experiences [11]. To improve health equity and minimize health disparities, healthcare systems are applying frameworks for addressing SDOH, engaging in community collaboration and cross-sector partnerships, and using technology for interventions [11‐14]. However, workforce readiness to address SDOH is a primary driver of program success and better patient outcomes [15]. To address workforce readiness, there is a need for a scale to measure provider SDOH competence. Currently, no known scales assess health service professionals’ competence to address SDOH in practice. To date there is an overreliance on qualitative assessments of SDOH knowledge, however, these studies are not generalizable to a larger population of health service professionals [16, 17]. The closest available Likert scale was developed to assess knowledge of the SDOH framework and understanding in five areas of empirical interests: awareness, understanding, and attitudes toward the SDOH framework [18]. The scale was a part of a larger project to evaluate health and social service systems however, the scale could not be validated or located in its entirety [18]. There were no other available Likert scales to assess SDOH competence for health service professionals known to the author at the time of this study’s development in 2019.

Competence is related to education and learning, therefore, to understand health service professionals’ SDOH competence the scale developed is based on a model of interprofessional education created by the Institute of Medicine (IOM) committee on measuring the impact of interprofessional education on collaborative practice and patient outcomes [19]. In addition, because addressing SDOH is aligned with cultural competence and social justice; the awareness, knowledge, and skills structure of the Multicultural Counseling Competencies and Standards was also utilized [20, 21]. The IOM model has four interrelated components that relate to overall health and system outcomes. The four interrelated components include (1) a learning continuum; (2) the outcomes of learning; (3) individual and population health outcomes; (4) system outcomes such as organizational changes, system efficiencies, and cost-effectiveness; and the major enabling and interfering factors that influence implementation and overall outcomes [19]. The ACN: SDH focuses on the outcomes of learning, in the model which includes attitudes/perceptions, knowledge/skills, collaborative behavior, and performance in practice [22]. The learning continuum (i.e., foundation and graduate education, and continuing professional development) should inform learning outcomes and impact the health care system along with patients; of specific interest to the current researcher is impacting individual health and population/public health. A key component to impacting individual health is understanding the gaps in learning outcomes for health service professionals’ (i.e., SDOH competence; IOM, 2015, p. 28) [15].

The purpose of the present studies was to develop and validate an SDOH competency assessment for health service professionals. Competence is related to education and learning, and the scale is developed based on a model of interprofessional education created by the Institute of Medicine (IOM) [15] along with the Multicultural Counseling Competencies and Standards [23, 24]. The research questions include: (1) What is the factor structure and reliability of the ACN: SDH? (2) Is the emergent factor structure of the ACN: SDH confirmed in a second sample?

Data collection began after IRB approval in Fall of 2019 for Study 1(EFA) and Spring of 2020 for Study 2 (CFA); these were two separate recruiting events. Participants were recruited using an email list of health service professionals and through health professional organization listservs. The following inclusion criteria were used to recruit and select participants for the study: (1) Foundation education in a healthcare field, (2) Professional identity as a healthcare provider, (3) Experience with healthcare practice as a provider, and (4) over 18. The informed consent from study participants was written. The survey started with a consent form that described the purpose of the study and associated risk, followed by the demographic questionnaire and the scales.

Best practice guidelines for scale development were followed, including developing an initial set of items and checking face validity using an expert panel [25]. An initial set of 163 items was created based on the theoretical foundation, recurring themes in the literature, and expert input. The scale utilized a five-point Likert-type scale, ranging from 5 (strongly agree) to 1 (strongly disagree).

Study 1 was completed by 220 participants; full results are in Table 2. In summary, the largest responses were in the following categories: Females 78.6% (n = 173), White 66.4% (n = 146), residing in the Southeast region of the United States 29.1% (n = 64), and identifying as mental/behavioral health service professionals 63.2% (n = 139). Study 2 was completed by 303 participants; full results are in Table 1. In summary, the largest responses were in the following categories: Females 80.5% (n = 244), White 52.8% (n = 160), residing in the Southeast region of the United States 40.9% (n = 124), and primarily identifying as a mental/behavioral health provider 50.2% (n = 152).

The 45 items of the ACN: SDH scale were subjected to principal components analysis (PCA), and variable reduction technique, using SPSS version 18. Prior to performing the PCA, the suitability of the data for factor analysis was assessed. Confirmatory factor analysis (CFA), a statistical procedure to test how well the measured variables represent the constructs, was conducted using maximum likelihood estimation as implemented in SPSS 28 AMOS. Maximum likelihood with robust standard errors (MLR) was used to estimate model parameters and the goodness-of-fit of all the CFA models was examined with: The root-mean-square-error of approximation (RMSEA) ≤ 0.06 (90% CI ≤ 0.06), standardized root mean square (SRMR) ≤ 0.08, comparative fit index (CFI) ≥ 0.90, and tucker-lewis index (TLI) ≥ 0.90 [27, 28]. Additionally, the chi-square/df ratio ≤ 3 rule was also used [29].

The suitability of the data was assessed through the following: Inspection of the correlation matrix revealed the presence of many coefficients of 0.30 and above. The Kaiser–Meyer–Olkin value was 0.92, exceeding the recommended value of 0.60 [30] and Bartlett’s Test of Sphericity [31] reached statistical significance, supporting the factorability of the correlation matrix (t (χ2 (990) = 6960.55, p < 0.001). The data was suitable for factor analysis.

Principal components analysis (PCA), initially revealed the presence of 9 components with eigenvalues exceeding 1, explaining a total of 69% of the variance. Inspection of the scree plot revealed the presence of six components; however, a model with four and five components was examined using varimax and oblimin rotations of the factor loading matrix. The model with six components explained a total of 61% of the variance, and it was preferred because of: (a) theoretical relevance and (b) eigenvalues on the scree plot leveled off after six. The varimax solution was used in subsequent analyses because the component transformation matrix had small values (0.22) indicating that components are not highly correlated which is the underlying hypothesis of the varimax solution. Lastly, a total of 20 items were eliminated because they either did not contribute to a simple factor structure and failed to meet a minimum criterion of having primary factor loadings of 0.4 or above, along with no cross-loadings of 0.3 or above.

Finally, a PCA, using varimax rotation was used on the 25-item ACN: SDH scale constrained, and a 6-factor solution was revealed, explaining 61% of the variance. Factor 1 contained six items that represented Action Toward Addressing SDOH and accounted for 33.96% of the variance. Factor 2 contained six items that represented SDOH Knowledge and accounted for 8.01% of the variance. Factor 3 contained three items that represented Negative Attitudes toward Addressing SDOH and accounted for 5.86% of the variance. Factor 4 contained three items that represented Systemic Accountability and accounted for 5.35% of the variance. Factor 5 contained four items that represented the Perception of the cause of SDOH and accounted for 4.04% of the variance. Factor 6 contained three items that represented School Preparation and accounted for 4.19% of the variance.

The Cronbach alpha reliability estimate for the 25-item total scale was 0.809. Item total correlations ranged from r = -0.28 to r = 0.88 with a mean item-total correlation of r = 0.25. The six subscales demonstrated the following satisfactory internal consistency estimates: Factor 1, (a = 0.84); Factor 2, (a = 0.92); Factor 3, (a = 0.74); Factor 4, (a = 0.79); Factor 5, (a = 0.87); and Factor 6, (a = 0.37). The study 1 results showed that the 25-item SDH: ACN has six dimensions which are concordant with the theoretical model.

The purpose of Study 2 was to confirm the dimensionality and evaluate the nomological network of the ACN: SDH. Table 3 shows the CFA model comparisons; in the first model, all 25 items were allowed to load on a single factor (SDH provider competence). The second model is a first order orthogonal model where all items are loaded on the defined dimensions from the EFA (i.e., the six factors) and these dimensions were not correlated. The third model is the same as the second model except the six factors were allowed to correlate (i.e., oblique). The fourth model is a second order model in which all items were allowed to load on their defined dimensions and all dimensions loaded on a second order factor of ‘SDH provider competence’. All 4 models showed poor model fit; three indicators were identified as having a modification index larger than 20 causing the poor model fit. The items deleted included AP3, AP4, and ATT9. The fifth model is a first order oblique model where all items are loaded on the defined dimensions (i.e., six factors); this model showed acceptable model fit. The fit indices for each model are presented in Table 4.

Model 5 first order oblique 6-factor model showed a better fit than the other models. Fitting the oblique model indicates that specific domains of provider SDOH competency are not interrelated. The chi square, CFI, RMSEA, and SRMR values indicate good to reasonable fit of model 5 (χ2 = 460.281, df = 194, CFI = 0.937, RMSEA = 0.067, SRMR = 0.048). The factor loadings for the preferred first order model (i.e., model 5) are presented in Table 4.

Internal consistency estimates of the ACN: SDH total score and subscale scores for Study 2 were high. The Cronbach alpha reliability estimate for the 22-item total score was 0.867. Item total correlations ranged from r = -0.27 to r = 0.89 with a mean item-total correlation of r = 0.23. The six subscales demonstrated satisfactory internal consistency estimates, listed in Table 4: Factor 1, (a = 0.85); Factor 2, (a = 0.94); Factor 3, (a = 0.79); Factor 4, (a = 0.81); Factor 5, (a = 0.86); and Factor 6, (a = 0.94). The total 22 item scale demonstrated satisfactory internal consistency. Study 2 results indicate that the 22-item ACN: SDH has six dimensions which is concordant with the theoretical model.

Pearson’s correlations between the ACN: SDH and the EMCC scales are summarized in Table 5. The ACN: SDH total and six subscale scores showed a low to moderate positive correlation to all subscales that are conceptually similar in content, which shows convergent validity. Discriminant validity is indicated by the inverse relationship between the EMCC and the ACN: SDH Perception subscale’ which is hypothesized to be conceptually different.

Subgroup analysis was conducted using ANOVA. A one-way between group analysis of variance was conducted to explore ‘healthcare provider industry’ on levels of SDH competency, as measured by the ACN: SDH. Participants were divided into four groups according to their healthcare provider industry (Group 1: Behavioral/Mental health, n = 152; Group 2: Medical health services, n = 93; Group 3: Educators, n = 39; Group 4: Other, n = 18). There was a statistically significant difference at the p < 0.01 level in ACN: SDH scores for the four groups: F (3, 301) = 6.499, p = 0.001. The effect size calculated using eta squared, was 0.06, which is a medium effect size [32]. Post-hoc comparisons using Tukey HSD test indicated that the mean score for the Behavioral/Mental health group (M = 80.03, SD = 10.80) was statistically different from that of the Medical Health Services (M = 76.09, SD = 11.49) and Other group (M = 69.11, SD = 14.04). The Education group (M = 76.67, SD = 10.40) did not differ significantly from either group.