The cultural adaptation and psychometric evaluation of the Mothers on Respect Index for Native Hawaiians and Pacific Islanders

Abstract

Background

Native Hawaiian and Pacific Islanders (NHPI) are disproportionately burdened by pregnancy-related deaths in the United States and have the lowest engagement in prenatal care compared to all other US racial groups. Aside from access barriers, studies suggest that NHPI face challenges with patient-clinician communication, perceived discrimination, and cultural conflicts within healthcare settings. This paper describes the cultural adaptation of the 14-item Mothers On Respect index for NHPI, originally developed by Vedam et al. (2017) for diverse communities in British Columbia, Canada, and reports the findings of the preliminary psychometric assessment of the adapted measure.

Methods

Data from 26 interviews with NHPI women, expert, and cognitive interviews were conducted to inform the adaptation. An online survey was administered to a sample of 90 NHPI women to assess construct validity, convergent validity, and internal reliability of the adapted measure using exploratory and confirmatory factor analyses.

Results

The adaptation resulted in substantial changes to the original measure, mainly by the addition of items related to ‘feeling cared for by and connected to the provider’ and ‘perceived threats hindering communication.’ The psychometric analyses identified a three-factor structure for the culturally adapted index and confirmatory factor analyses were employed to refine the measure. The result was a 25-item index with acceptable goodness of fit indices, high internal reliability (Cronbach’s alpha of 0.96, 95% CI = .94-.97) and convergent validity with a related scale. Overall, participants in this sample indicated high levels of respectful care; however, people who received < 8 prenatal care visits had significantly lower ratings on average.

Conclusions

Our findings suggest that the elements valued by NHPI are not fully captured in existing measures of respectful maternity care. Efforts to assess more discrete aspects of the patient-provider relationship for culturally distinct and racialized groups could help improve the quality of care and advance equity in maternal and perinatal health marginalized communities.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-024-06856-7.

Introduction

In 2023, the Center for Disease Control and Prevention (CDC) reported, for the first time, disaggregated the data of Non-Hispanic Native Hawaiians and Pacific Islanders (NHPI) from Non-Hispanic Asian Americans and found that in 2017–2019, NHPI had the highest ratio (6.8 per 100,000 live births) of pregnancy-related deaths in the United States (U.S.) compared to Hispanics and Non-Hispanic Blacks, American Indians/Alaskan Natives, Whites, and Asians [1]. The CDC report is notable because previous reports and multiple research studies examining maternal health outcomes have rendered invisible NHPIs’ experiences due to aggregation of data with Asian American populations or other groups [2, 3]. This has contributed to the underrepresentation of NHPI in maternal and perinatal health research [4]. There is also an absence of culturally-adapted measures of maternal and perinatal health experience for NHPI communities to ensure their valid assessment.

In the U.S., pregnancy-related morbidity and mortality are on the rise [2, 5] and there are stark differences in risk across racial and ethnic groups commonly identified from federal data. The observed disparity in the maternal mortality rates for NHPI adds to the increasing and concerning evidence of racial inequities in maternal and perinatal health, particularly among Black and American Indian/Alaska Native pregnant and birthing people. The specific causal pathways of racial disparities are complex, interrelated and multilevel, but studies increasingly highlight the role of societal inequities as upstream drivers of risk for Black, Native American and NHPI groups [6–11]. These broad societal inequities stem from structural racism and impact neighborhood environments, economic wellbeing and stability, and health care system factors that include issues of access and quality of care [6, 9, 12]. Frameworks developed by prominent authors in the field propose the various forms of racism interconnect and are perpetuated contemporarily, manifesting in multiple ways – including exposure harms that increase health risk and via biological mechanisms [6, 9]. While, in studies investigating disparities between Black and White populations, individual-level factors like hypertension have been identified as significant drivers of maternal health, these individual-level factors account for only a portion of the differences in risk [13].

Revealing the specific mechanisms producing the racial disparities in maternal and perinatal health across different racial groups in the U.S. continues to be a focus of research. One factor that stands out when trying to understand disparities facing NHPI in the U.S. is that NHPI have lower engagement in prenatal care compared to other racial/ethnic groups [14]. Specifically, NHPI have the highest “late to no care” prenatal care initiation (20.2%) compared to Whites (4.5%), Blacks (9.9%), American Indians or Alaska Natives (13.1%), Asians (4.9%), and Hispanics (7.7%) [15]. Given the substantial burden in pregnancy-related deaths borne by NHPI, and the pattern of low engagement in prenatal care, identifying the underlying reasons for this disparity is a public health imperative.

For NHPI communities, research to understand the factors contributing to low levels of engagement with prenatal care could reveal strategies for improving maternal health outcomes. There is evidence that NHPI communities experience racial discrimination in healthcare settings and report negative perceptions of pregnancy-related healthcare, like what has been expressed by other racialized communities [16–19]. Communication, perceived discrimination, conflicts between the healthcare system and NHPI cultural norms, and negative attitudes and treatment encountered in healthcare settings are all identified obstacles to NHPI receiving quality care [11, 20–24]. In a study examining obstetrician-gynecologists’ (OB-GYNs) perspectives on their experiences providing labor and delivery care for Pacific Islanders, it was noted that communication barriers might hinder shared decision-making [22]. Communication barriers included challenges with language interpretation services, as well as differences in communication styles. Feeling rushed during short prenatal visits and feeling discriminated against were also identified in the literature as barriers to quality care [11]. It is important to note that these studies conducted in various states (Utah, Arkansas, Hawai‘i, and Guam) with different Pacific Islander communities are highlighting communication, attitudes and interpersonal treatment challenges in the context of care.

Although the importance of patient-clinician communication and patient-centered care has been promoted by national medical and public health agencies, such as the Institute of Medicine, Center for Medicare and Medicaid Services, The Joint Commission, Agency for Healthcare Quality and Research, there are few validated instruments to evaluate patient-centered care [25] or the patient-clinician relationship in the context of care for pregnant people [26]. None, to our knowledge, have been developed or validated for NHPI. This paper describes the methods and outcomes of a study that aimed to culturally adapt the Mothers on Respect (MOR) index to evaluate critical aspects of patient-clinician relationships and communication for NHPI in the context of maternity care. The MOR Index was created by investigators of the Birth Place Lab in British Columbia using a community-based participatory research approach [27]. It measures the perceived quality of maternal care in reference to the patient-clinician relationship.

Methods

Study overview

This study involved the cultural adaptation and evaluation of the MOR index, a measure of respectful care for pregnant and birthing individuals. Additional details regarding this measure are described under the section titled, “Phase 1: Cultural adaptation of the Mothers on Respect Index…” Fig. 1 depicts the study design and the two phases of this study. Three community-based organizations (CBOs) collaborated on this study and assisted with outreach and recruitment of participants. The CBOs are: (1) We Are Oceania, a non-profit public charity serving Pacific Islanders throughout Hawai‘i, (2) Kula no nā Po ‘e Hawai‘i, a service organization serving Native Hawaiian Homestead communities, and (3) Healthy Mothers Healthy Babies Coalition of Hawaiʻi, a non-profit network of organizations supporting pregnant and parenting people in Hawai‘i. All three organizations have offices on the island of Oahu, but have networks that extend throughout the entire state. At the time of the study none of the CBOs operated as a healthcare facility, yet each CBO is involved in supporting the health and well-being of their NHPI clients. For example, programs offered by these CBOs include helping community members access information about health insurance, social supports, and vaccinations in their native languages, organizing food distribution to families in need during the COVID-19 pandemic, and providing needed referrals and supportive services to pregnant women. Working with partners that did not provide clinical care was a strategic decision by the study team to reduce the possibility of participants feeling conflicted about sharing honestly about their maternity care experiences with providers and clinics that they utilize.

Fig. 1.

Study flow and details of modifications at each step/phase. Note. MORi v#x = Culturally adapted Mothers on Respect index, Original version #, Modified version x. NHPI = Native Hawaiians and other Pacific Islander

The study was conducted from September 2020 through April 2022 and was approved by the Institutional Review Board of the University of Hawai‘i. Informed consent to participate was obtained from all participants in the study prior to data collection. The following sections will detail the methods and results of the cultural adaptation (Phase 1) and the psychometric analysis (Phase 2) of the adapted measure.

Phase 1: Cultural adaptation of Mothers on Respect (MOR) index for NHPI communities

The cultural adaptation process involved three steps: 1) key informant interviews, 2) expert panel review, and 3) cognitive interviews (see Fig. 1) [28]. The MOR index is a survey tool comprised of 14 items using a 6-point Likert response scale, scored from one = strongly disagree to six = strongly agree. The instrument was selected for cultural adaptation for several reasons: 1) it allows for assessment of prenatal and/or birth experiences; 2) it includes seven items used in a cross-sectional, nationally drawn survey conducted in the U.S., enabling the possibility of comparison of individual items across multiple samples; and 3) it is a relatively concise measure addressing factors (patient-provider communication and discrimination) identified in studies of NHPI groups as barriers to care or disruptions in the patient-provider encounter. The MOR index was evaluated in Canada and in the U.S. as part of the Giving Voice to Mothers study [27, 29]. Respondents included women from communities of color and women who choose home births (N = 2,357). A small percentage (3.4%) of U.S. sample identified as either Native Alaskan or NHPI, but their data were aggregated into a single racial/ethnic category termed “Indigenous” in the psychometric analysis. Thus, the tool has not been assessed specifically with an NHPI sample and it is unclear that the scale adequately captures the concept of “respect” as perceived and conceptualized by NHPI groups.

Key informant interview methods

The purpose of the key informant interviews was to better understand the maternity experiences of NHPI women and to collect feedback on the original MOR Index items. Key informant interviews were conducted via video teleconference due to restrictions in place due to risks related to COVID-19. When the pandemic restrictions eliminated the ability to conduct interviews face-to-face with participants and alternatives needed to be identified, the CBOs were consulted for their expertise engaging with NHPI community members. NHPI serving community partners helped develop recruitment strategies and recruited participants for the project. The enrollment target was 40 individuals or until theme saturation was reached. Eligibility criteria for participants included being 18 years or older NHPI woman, having delivered no more than two years prior in the state of Hawai‘i, comprehending English, and having access to the internet. Interview participants were provided a $40 electronic gift card for participating.

The semi-structured interviews lasted approximately two hours and were organized into two parts. The first part of the interview involved experiences and perspectives of the participants’ delivery. The second half was feedback on the original MOR Index. Qualitative content analysis was used to iteratively develop a coding frame and to generate themes [30]. Coding was conducted by two coders (RD,AC). Themes were derived from assessing common concepts, experiences, and associations. Results from the themes identified through analysis in the first part of the interview and feedback specific to the original version of the MOR index (Part 2 of the interview) were utilized to inform the first round of modifications (MORi v2a) to the original MOR index. For example, if three or more participants had similar comments about a particular question, that aspect was addressed through a modification or addition. No questions were recommended to be removed by three or more participants.

Expert panel review methods

The next step was to share the modified version (MORi v2a) with experts to review. Experts were defined as researchers and clinicians with expertise in survey administration related to health or healthcare research and/or experience working with NHPI communities. A list of experts was developed by the study team and those individuals were recruited individually via email. If they accepted the invitation, an assessment form was given to the Expert Reviewer to assess content validity in four areas: relevance, appropriateness, clarity, and specificity [31]. Each item was evaluated on a 4-point scale, e.g., one (the item is not relevant) to four (the item is highly relevant) [31]. The assessment materials provided to the reviewers contained the modified MOR Index (MORi v2a) and the content validity scale (see Additional file 1). If the expert reviewer rated any item 2 or lower, they were asked to justify their rating with a brief text entry. Additionally, there were two text entry questions asking respondents to provide additional feedback and comment on the overall representation of respectful maternity care. Content validity scores were calculated based on the procedures outlined in Yusoff (2019) [31]. Based on scores and text responses the index (MORi v2a) was then revised a second time, resulting in MOR index version 2b (MORi v2b).

Cognitive interview methods

The final step in the cultural adaptation was to conduct cognitive interviews. The eligibility criteria for cognitive interview participants, the outreach, and the recruitment were the same as in Step 1 for the key informant interviews. A $10 electronic gift card was provided at the end of the interview. During the cognitive interview, participants were asked about their comprehension/interpretation of each item in the adapted MOR index. The responses were analyzed for themes or patterns related to problems with items or contextual differences. The revisions resulted in a third version of the index (MORi v2c).

Phase 2: Psychometric evaluation methods

The 32-item culturally adapted MOR index (MORi v2c) with response options on a 6-point Likert scale was psychometrically evaluated for construct and convergence validity. Higher scores on the MOR index indicate a better assessment of experiences. The Person-Centered Prenatal Care (PCPC) scale [26] was incorporated to examine the convergent validity of the culturally adapted MOR index. PCPC is a reliable and validated tool for assessing the patient's experience during pregnancy. It consists of 26 Likert 4-point items, with scores ranging from 0 (none of them) to 4 (all of them). The Cronbach’s alpha of the PCPC for this study was 0.95, demonstrating excellent internal consistency.

The culturally adapted MOR index (MORi v2c) was embedded into an online survey that included demographic questions, healthcare facility-level questions, and questions regarding the pregnancy and delivery. The survey was created on the Research Electronic Database Capture (REDCap) platform [32]. The demographic factors included age, insurance, race/ethnicity, marital status, education level, number of people in the household, household income, and delivery year (pre-pandemic, pandemic between March 2020 and March 2021, and post-pandemic). Examples of broader contextual factors examined include: type of insurance coverage, type of delivery, and number of prior births.

Recruitment for the online survey

Participants were recruited through flyers, listservs from collaborating CBOs, social media, and community events. The eligibility criteria encompassed women who: 1) were over 18 years old, 2) self-identified as NHPI, 3) had given birth within the past three years in Hawai‘i, and 4) were proficient in reading English. A $10 electronic gift card was provided for anyone who opted to provide an email address upon survey completion.

Statistical methods for the online survey

Before conducting psychometric analysis, we performed descriptive analysis to elucidate the characteristics of the study participants. To evaluate the validity of the adapted MOR, both construct validity and criterion validity were employed. Because the key informant interviews suggested multiple additions to the scale that increased the number of items substantially, the research team decided to treat the culturally adapted MORi as new scale. Construct validity underwent scrutiny through exploratory factor analysis (EFA), followed by confirmatory factor analyses (CFAs). EFA, employing Kaiser’s rule (eigenvalue > 1) and scree plot rule, determined the retention of factors. Given the theoretical interrelatedness of domains, we opted for promax oblique rotations, allowing for correlation between rotated factors. In addition, we excluded items with loadings below 0.3 or those that loaded onto multiple factors. Despite the challenge of a small participant pool, we conducted CFAs for each identified factor to refine items. Model fit was evaluated using the Satorra-Bentler scaled chi-square test, root mean square error of approximation (RMSEA) and its 90% confidence interval (CI), standardized root mean square error (SRMR), comparative fit index (CFI), and Tucker-Lewis index (TLI). A good fit was indicated by a non-significant chi-square test, RMSEA < 0.05, SRMR < 0.06, CFI > 0.95, and TLI > 0.95. Models were adjusted to meet these criteria, guided by modification indices, theoretical considerations and priorities identified in key informant interviews.

Construct validity was assessed through hypothesis testing with prenatal care-related characteristics and correlation analysis with PCPC. One-way analyses of variance were employed to explore associations between confirmed factors of the adapted MOR and prenatal care-related characteristics, such as the number of visits with the provider they saw most often for prenatal care. Tukey’s post-hoc test was conducted to explore for significant association. Convergent validity was evaluated through Pearson’s correlations between the adapted MOR and PCPC. Reliability for the PCPC and the adapted MOR and its subscales were assessed using Cronbach’s alpha. A p-value less than 0.05 was considered statistically significant, and all analyses were implemented using SAS software version 9.4 [33].

Results

Phase 1 – Results of the cultural adaptation (see Fig. 1)

Key informant interviews

Twenty-six NHPI key informants were enrolled and interviewed. Half of the participants identified as Native Hawaiian. Based on participant feedback on the MOR index, the number of items responding to the prompt, “During my pregnancy, I held back from asking questions or discussing my concerns because of” was expanded. For example, participants expressed their concerns about being judged or their provider not taking their concerns seriously [19]. Additionally, items reflecting themes of communication, caring, and connection were added to the scale based on themes emerging from the interviews (e.g., “Overall, during my prenatal visits I felt my provider cared about me and my family”) [19]. There were 31 changes made at this stage, including the addition of 23 items resulting in a 37-item scale (MORi v2a).

Expert reviews

Nine experts were enrolled for expert review. Overall, expert reviewers determined high relevance (0.98 out of 1.0) and appropriateness (0.97) through content validity scoring. Clarity and specificity were scored slightly lower (0.91, 0.89 respectively). Determinations about retaining or changing items were made considering scores and qualitative analyses from expert comments and key informant interviews. Revisions made in the expert review stage were largely focused on items that were added in the first step. Six items were removed and multiple wording changes were made resulting in a 32-item version of the adapted scale (MORi v2b).

Cognitive interviews

Five cognitive interviews were conducted. Participants demonstrated good comprehension and interpretation of the revised index. One participant concluded, “Seems like you are asking what kind of relationship I had with my provider,” providing a strong indication of clarity and understanding. The cognitive interviews resulted in a limited number of very minor changes to the tool. However, an option to skip a question regarding “friends and family” was added due to the Coronavirus Pandemic, which due to social distancing policies could have prevented friends and family from attending appointments.

The original MOR index comprised 14 items, and after culturally adapting the measure to NHPI, the adapted version after Phase 1 (MORi v2c) comprised 32 items (see Fig. 1). Most of the index changes occurred in Steps 1 and 2 of Phase 1. This 32-item version (MORi v2c) was then tested for validity in Phase 2, described below.

Phase 2—Results of the Psychometric Evaluation

Participant characteristics

Ninety people participated in the online survey. Table 1 illustrates the characteristics of the participants. Most respondents identified as Native Hawaiian (80.0%) and fell within the 25–34 age group (64.4%). Half of the participants had completed at least a bachelor’s degree. Sixty percent were covered by Medicaid insurance and just under half (47.8%) were reporting on their first birth. Approximately three-quarters of the respondents (75.6%) had an initial prenatal visit between 6–12 weeks, and 55.6% reported having at least 10 prenatal care visits. Most participants saw a “doctor” most often (82.2%) for prenatal care. The next most common type of provider seen most often was a certified nurse midwife (11.1%).

Table 1.

Online survey participant characteristics

Characteristics	n	%
Maternal demographics
Age (years)
20–24	18	20.0
25–34	58	64.4
36–45	14	15.6
Race (primary)a
Native Hawaiian	72	80.0
Pacific Islander	11	12.2
Other	7	7.8
Hispanic or Latino	8	8.9
Marital status
Married/Living with a partner	67	74.4
Single	23	25.6
Educational level
High school or less/Some college	33	40.0
Associate degree or other specialized training	9	10.0
Completed bachelor’s degree	21	23.3
Completed graduate degree or more	24	26.7
Total yearly household income
Less than $30,000	14	16.1
$30,000-$49,999	14	16.1
$50,000-$99,999	3	36.8
$100,000-$200,000	22	25.3
More than $200,000	5	5.7
Type of insuranceb
Medicaid	54	60.0
Private insurance	46	51.1
Other	4	4.4
Obstetric factors
Number of times given birth
1	43	47.8
2–3	33	36.7
≥ 4	14	15.6
Number of weeks at initial prenatal visit
≤ 5	18	20.0
6–12	68	75.6
13–28	4	4.4
Total prenatal care visits
≤ 7	21	23.3
8–9	19	21.1
≥ 10	50	55.6
Most frequent facility for prenatal care
Private	50	55.6
Community health center	12	13.3
Integrated healthcare system	20	22.2
Hospital clinic	8	8.9
Type of Delivery
Vaginal (any)	79	87.8
Cesarean	5	5.6
Cesarean (unplanned/emergency)	6	6.7
Had health problems during pregnancy	26	28.9
Had health problems during delivery	24	26.7
Provider and hospital setting
Type of provider seen most often
Doctor	74	82.2
Certified nurse-midwife	11	12.2
Other midwife	5	5.6
Number of visits with provider seen most often
3–5 times	15	16.7
6–9 times	39	43.3
10 or more times	36	40.0
Delivered by provider seen most often
Yes	58	64.4
No	32	35.6
Delivered pre-pandemicc
Yes	20	22.2
No	70	77.8
Continuous Variables	M	SD
Number of people in household	5.2	2.7
Years living in Hawai`i	25.9	7.0

Note. N = 90. M and SD are used to represent mean and standard deviation, respectively. ^aParticipants who identified with multiple race categories were asked to self-identify a “primary” race. The “Other” category includes people who identify as Native Hawaiian or Pacific Islander but selected another race as their “primary” race. ^bParticipants can have more than one insurance. ^c “Delivered pre-pandemic” was defined as Yes if a participant delivered a baby before March 2020

Factor analyses

One item was excluded from factor analyses due to a 12.2% non-response rate among participants. The initial EFA revealed 3 factors for the adapted MOR index (see Table 2). During the EFA, four items were deleted: three displayed loadings on multiple factors, while one failed to load onto any of the 3 factors. These factors are labeled as follows: 1) dignity, caring, and connection (DCC), 2) perceived threats (PT), and 3) perceived discrimination (PD).

Table 2.

Results from explanatory factor analysis on Adapted Mothers on Respect Index (MORi v2c)

Item	M	SD	Factor Loading			h2
			1	2	3
1. I felt comfortable asking my provider questions.	5.38	1.00	.50	.27	.09	.55
2. I felt comfortable answering questions my provider asked me.	5.59	0.67	.19	.54	.19	.58
3. My provider clearly explained things to me.	5.30	0.94	.85	.03	-.01	.75
4. My provider listened carefully to what I said.	5.32	0.97	.99	-.04	-.11	.88
5. My provider answered my questions.	5.42	0.83	.91	-.04	-.15	.73
6. I felt comfortable saying “no” to health care options offered to me.	5.04	1.24	.00	.82	-.11	.62
7. I felt comfortable saying “yes” to health care options offered to me.	5.38	0.91	.19	.74	-.17	.65
8. I felt pressured by my provider to follow their recommendations. ®	4.30	1.66	.20	.05	.17	.11
9. My provider helped and supported me when I needed it.	5.39	0.83	.88	.02	.02	.79
10. My provider comforted and encouraged me when I needed it.	5.36	0.93	.87	-.01	.10	.82
11. My provider treated me with respect.	5.59	0.67	.78	.11	-.10	.67
12. My provider treated me family and friends with respect.α	5.53	0.73	-	-	-	-
13. I felt my provider cared about me.	5.40	0.88	.94	-.18	.14	.80
14. I knew my provider well.	4.78	1.30	.60	.26	.04	.65
15. I felt a personal connection with my provider.	4.70	1.39	.74	.10	.04	.67
16. My provider respected my personal and cultural preferences.	5.33	0.86	.61	.25	.05	.64
17. There were times that I felt I was treated poorly by my provider because of my race/ethnicity or cultural background. ®	5.44	0.98	.16	.00	.83	.80
18. There were times that I felt I was treated poorly by my provider because of my language. ®	5.54	0.90	-.05	-.03	.99	.92
19. There were times that I felt I was treated poorly by my provider because of my education level. ®	5.56	0.84	.00	-.04	.99	.94
20. There were times that I felt I was treated poorly by my provider because of my sexual and/or gender orientation. ®	5.59	0.75	.02	-.15	1.02	.94
21. There were times that I felt I was treated poorly by my provider because of my weight. ®	5.28	1.13	-.06	.23	.65	.55
22. There were times that I felt I was treated poorly by my provider because of my health insurance. ®	5.50	0.96	-.05	.11	.90	.86
23. During conversation with my provider I held back from asking questions or discussing my concerns because my provider seem too busy. ®	4.77	1.61	.56	.36	-.01	.68
24. During conversation with my provider I held back from asking questions or discussing my concerns because my provider might think I was being difficult. ®	5.08	1.34	.43	.51	.02	.72
25. During conversation with my provider I held back from asking questions or discussing my concerns because I thought I would be judged. ®	5.01	1.46	-.02	.89	.03	.80
26. During conversation with my provider I held back from asking questions or discussing my concerns because I thought my questions or concerns would not be taken seriously. ®	5.13	1.26	.26	.72	-.01	.80
27. During conversation with my provider I held back from asking questions or discussing my concerns because I thought that my provider did not like me. ®	5.51	0.72	.25	.51	.21	.63
28. During conversation with my provider I held back from asking questions or discussing my concerns because I didn’t know what to say or how to ask my questions. ®	5.01	1.33	.00	.80	-.06	.60
29. During conversation with my provider I held back from asking questions or discussing my concerns because I did not feel it was my place to ask questions about the care I was receiving. ®	5.39	0.97	-.08	.93	.12	.89
30. During conversation with my provider I held back from asking questions or discussing my concerns because I was scared or intimidated. ®	5.31	1.07	-.14	1.01	-.01	.87
31. During conversation with my provider I held back from asking questions or discussing my concerns because I was in a rush to end the conversation. ®	5.27	1.11	.16	.69	.08	.68
32. During conversation with my provider I held back from asking questions or discussing my concerns because I did not believe my provider could answer my questions or address my concerns. ®	5.38	0.94	.37	.44	.16	.64

Note. All items were 6 type Likert-scales scored as 1 = strongly disagree ~ 6 = strongly agree. ® item was reversely scored. h² indicates communality which is the amount of variance in the item explained by the factors. ^αThis item was not included in the explanatory factor analysis because 11 participants (12.2%) did not respond to this question

Subsequent CFAs led to the removal of two additional items due to high correlations with other items (Additional file 2) and the addition of error covariances to meet the model fit statistics (see Table 3). For DCC, one item was removed (item 5) that was highly correlated with other items and added three error covariances (between items 4 and 11, items 4 and 15, and items 14 and 15), finalizing 10 items. For PT, one error covariance was added (between items 27 and 31), finalizing 10 items. For PD, one item (item 20) was removed and one error covariance (between items 17 and 21) was added. After these adjustments, the sum for the overall adapted MOR and each factor was computed. The means of the overall scores were 132.5 (SD = 18.3) for adapted MOR, 52.6 (SD = 8.1) for DCC, 52.6 (SD = 9.1) for PT, and 27.3 (SD = 4.3) for PD. All scored were skewed to the right, indicating that participants generally perceived a high level of respect.

Table 3.

Model fit statistics of confirmatory factor analyses for three factors of adapted Mothers On Respect Index

Factor	Process	X2	df	P	RMSEA (90% CI)	SRMR	CFI	TLI
Original	8 items out of 14 items	79.38	20	< .01	.18 (.14, .23)	.23	.69	.57
1: Dignity, Caring, & Connection	11 items	96.62	44	< .01	.11 (.18, .15)	.07	.90	.87
	Remove 1 item	66.04	35	< .01	.10 (.06, .14)	.05	.93	.91
	Add three error covariances	36.97	32	.25	.04 (.00, .09)	.04	.99	.98
2: Perceived Threat	10 items	54.58	35	.02	.08 (.03, .12)	.06	.94	.92
	Add an error covariance	42.00	34	.16	.05 (.00, .10)	.06	.97	.97
3: Perceived Discrimination	6 items	7.08	9	.63	.00 (.00, .10)	.05	1.00	1.04
	Remove 1 item	7.09	5	.21	.07 (.00, .17)	.04	.96	.93
	Add one error covariance	4.72	4	.32	.04 (.00, .17)	.02	.99	.97

Note. X² = Satorra-Bentler scaled chi-square test. df degrees of freedom of chi-square test, RMSEA root-mean square error of approximation, CI confidence interval, SRMR standardized root mean square residual, CFI comparative fit index, TLI Tucker Lewis index of factoring reliability

Internal consistency & convergent validity

Table 4 presents the loading scores of the identified factors. Cronbach's alphas demonstrated strong internal consistency, with values of 0.96 (95% CI = 0.94-0.97) for the overall adapted MOR, 0.95 (95% CI = 0.93-0.96) for DCC, 0.94 (95% CI = 0.92-0.96) for PT, and 0.94 (95% CI = 0.91-0.95) for PD. For assessing convergent validity using PCPC, Pearson’s correlations were as follows: 0.74 with the adapted MOR, 0.65 with DCC, 0.73 with PT, and 0.40 with PD (see Additional file 3).

Table 4.

Loading of final confirmatory factor analyses of adapted Mothers On Respect Index (MORi v3)

Item	Loading	Cronbach’s α (95% CI)
Factor 1: Dignity, Caring, & Connection (10 items)		.95 (.93, .96]
1. I felt comfortable asking my provider questions (Item 1)	.69
2. My provider clearly explained things to me (Item 3)	.83
3. My provider listened carefully to what I said (Item 4)	.90
4. My provider helped and supported me when I needed it (Item 9)	.93
5 My provider comforted and encouraged me when I needed it. (Item 10)	.90
6. My provider treated me with respect (Item 11)	.72
7. I felt my provider cared about me (Item 13)	.89
8. I felt a personal connection with my provider (Item 15)	.76
9. My provider respected my personal and cultural preferences (Item 16)	.83
10. During conversation with my provider I held back from asking questions or discussing my concerns because my provider seem too busy (Item 23)	.75
Factor 2: Perceived Threats (10 items)		.94 (.92, .96]
1. I felt comfortable answering questions my provider asked me (Item 2)	.72
2. I felt comfortable saying “no” to health care options offered to me (Item 6)	.72
3. I felt comfortable saying “yes” to health care options offered to me (Item 7)	.75
4. During conversation with my provider I held back from asking questions or discussing my concerns because I thought I would be judged (Item 25). ®	.88
5. During conversation with my provider I held back from asking questions or discussing my concerns because I thought my questions or concerns would not be taken seriously (Item 26). ®	.85
6. During conversation with my provider I held back from asking questions or discussing my concerns because I thought that my provider did not like me (Item 27). ®	.72
7. During conversation with my provider I held back from asking questions or discussing my concerns because I didn’t know what to say or how to ask my questions (Item 28). ®	.75
8. During conversation with my provider I held back from asking questions or discussing my concerns because I did not feel it was my place to ask questions about the care I was receiving (Item 29). ®	.95
9. During conversation with my provider I held back from asking questions or discussing my concerns because I was scared or intimidated (Item 30). ®	.93
10. During conversation with my provider I held back from asking questions or discussing my concerns because I was in a rush to end the conversation (Item 31). ®	.79
Factor 3: Perceived Discrimination (5 items)		.94 (.91,]95)
1. There were times that I felt I was treated poorly by my provider because of my race/ethnicity or cultural background (Item 17). ®	.85
2. There were times that I felt I was treated poorly by my provider because of my language (Item 18). ®	.96
3. There were times that I felt I was treated poorly by my provider because of my education level (Item 19). ®	.96
4. There were times that I felt I was treated poorly by my provider because of my weight (Item 21). ®	.65
5. There were times that I felt I was treated poorly by my provider because of my health insurance (Item 22). ®	.91

Note. N = 90. CI Confidence interval. All items are 6 type Likert-scales scored as 1 = strongly disagree ~ 6 = strongly agree. Reverse-scored iteams are denoted with an ®

Association between adapted MOR and other factors

Table 5 illustrates the relationships between the three confirmed factors of the adapted MOR and various prenatal care-related characteristics. Participants who primarily attended a community health center for their prenatal care had significantly lower scores on the adapted MOR index. This trend was observed for the overall score, DCC, and PD. In addition, lower scores on the adapted MOR index were significantly associated with having less than seven prenatal care visits. Fewer than seven prenatal care visits were linked to lower scores on DCC, while seeing the same provider for prenatal care 10 or more times was associated with higher DCC scores. Also, PT is significantly associated with number of times a participant had given birth.

Table 5.

Associations between Adapted Mothers on Respect (MORi v2c) and participant/healthcare characteristics

Variable	aMOR			DCC			PT			PD
	M	SD	Pa	M	SD	Pa	M	SD	Pa	M	SD	Pa
Number of times given birth			.04			.06			.01α,β			.99
1	127.6	20.4		50.5	8.9		49.8	10.6		27.3	4.4
2–3	135.7	15.3		53.7	7.2		54.6	6.3		27.4	3.0
≥ 4	140.2	14.0		55.9	7.2		56.9	6.0		27.4	6.6
Education level			.35			.12			.45			.95
High school or less/some college	135.9	14.2		54.7	6.0		54.1	7.7		27.2	4.7
Associate degree	129.8	25.0		50.0	11.2		52.7	11.6		27.1	4.3
Bachelor’s degree or higher	130.3	19.7		51.3	8.9		51.5	9.5		27.5	4.1
Total prenatal care visits			.02α,β			.02α,β			.11			.13
≤ 7	123.0	19.8		48.0	9.2		49.1	9.9		25.9	4.8
8–9	137.2	17.3		54.3	7.3		54.3	9.6		28.6	2.3
≥ 10	134.7	16.9		53.8	7.6		53.5	8.1		27.4	4.6
Most frequent facility for prenatal care			< .01α,β			< .01α,β,γ			.13			< .01α,δ,ε
Community health center	116.2	19.2		43.3	9.4		48.2	9.5		24.7	3.4
Private	134.0	16.3		53.3	6.8		52.5	8.8		28.3	3.3
Integrated healthcare system	139.5	16.6		55.6	8.1		55.9	7.1		28.1	3.0
Other	130.3	21.1		54.3	6.6		52.4	11.5		23.6	9.1
Number of visits with provider seen most often			.05			< .01β,ζ			.15			.70
3–5 times	126.5	20.7		48.3	10.3		51.7	10.1		26.5	4.9
6–9 times	129.7	18.5		51.2	8.2		51.0	9.2		27.5	3.1
10 or more times	138.1	15.9		55.8	6.2		54.9	7.9		27.4	5.2
Delivered pre-pandemicb			> .99			.81			.89			.88
No	132.5	18.4		52.7	8.5		52.6	9.2		27.3	4.4
Yes	132.5	18.2		52.2	7.4		52.9	8.1		27.5	4.3
Type of insurancec
Medicaid			.67			.91			.63			.57
No	131.5	19.4		52.4	8.0		52.1	9.3		27.0	4.4
Yes	133.2	17.7		52.6	8.5		53.0	8.8		27.5	4.3

Note. N = 90. aMOR Adapted Mothers on Respect (possible range: 25 to 150). DCC Dignity, Caring, & Connection (possible range: 10 to 60). PT Perceived Threats (possible range: 10 to 60). PD Perceived Discrimination (possible range: 5 to 30). ^aOne-way analysis of variance was conducted on column variable. ^b “Delivered pre-pandemic” was defined as Yes if a participant delivered before March 2020. ^cParticipants can have more than one insurance

^αTukey’s post-hoc test reveals significant difference between the first and the second categories. ^βTukey’s post-hoc test reveals significant difference between the first and the third categories. ^γTukey’s post-hoc test reveals significant difference between community health center and other. ^δTukey’s post-hoc test reveals significant difference between private and other. ^εTukey’s post-hoc test reveals significant difference between integrated healthcare system and other. ^ζTukey’s post-hoc test reveals significant difference between 6–9 times and 10 or more times

Discussion

Through a community-engaged, multiphase process of cultural adaptation, we were able to adapt and evaluate the psychometric performance of a measure to better capture aspects of respectful care important to NHPI pregnant and birthing people. The original MOR index was revised with input from multiple sources and tested for its construct and convergent validity with the PCPC measure. The result is a 25-item 3-factor scale that differed from the original MOR index in two important ways. First, it expanded on specific threats perceived by patients concerning their interactions with maternity care providers/healthcare settings. Second, it expanded on the items focused on compassionate, connected, and respectful care demonstrated by maternity care providers.

There is scant research on culturally adapted measures for Native Hawaiian and Pacific Islander populations. A recent study out of Hawaiʻi adapted a Historical Loss Scale for Native Hawaiians that was originally validated among American Indian adults in the U.S. [34]. To our knowledge no measures related to maternity care experiences have been culturally adapted specifically for NHPI. However, a scoping review of validity studies of the Edinburgh Postnatal Depression Scale (EPDS) culturally adapted for various Southeast, and South Asian populations found that in addition to language, cultural nuances were cited as contributing to the lower cutoff scores among adapted compared to the original version which was validated in Scotland [35, 36]. For example, how people from different cultural groups expressed emotion, experienced emotion (e.g., somatization of postpartum depression), and societal stigma in the culture were noted [35]. While the focus of this study is quite different, the importance of cultural nuance in measure validation and adaptation is similar. The adaptation of the tool within this sample did not result in the addition of specific cultural practices. However, participants’ cultural perspectives came across in the context of values and expectations related to communication and the value placed on the expression of caring and dignity within the patient-clinician interaction.

While not as obvious as the incorporation of, for example, indigenous medicines or therapeutic massage, which remain important practices in many NHPI families, the results reveal the significance of relationships across Pacific cultures. By extension, the findings highlight how communication in the context of care may or may not reflect the importance of nurturing relationships. For example, new items focused on the feeling of being “cared” for, “supported,” and “encouraged” reflect the common Pasifika understanding of “relationship” – which is more accurately expressed in Pacific languages, for example the Tongan term “Vā” can be translated into English to mean the “space between.” This space represents an opportunity for connection that requires nurturing and reciprocity. This point is illustrated in a quote from the Samoan Poet, Albert Wendt (1996) who wrote, it is “not space that separates but space that relates, that holds separate entities and things together” [37].

The differences between the original and the adapted MOR index for NHPI emphasize the relationship and connection between the patient and clinician over aspects of autonomy. This shift away from Western notions of autonomy in the findings may reflect the relational orientation to health and the interdependent nature of family structures and decision-making processes within many NHPI communities [38–40]. The more muted emphasis on autonomy could also indicate that patients view providers as the authority within healthcare interactions. Some Pacific Islander key informants suggested that it is not uncommon for people in their communities to see a doctor as an “authority,” a role tied to cultural norms and expectations of behavior that demonstrate respect, but do not eliminate the possibility of shared decision-making. Less emphasis on autonomy should not be interpreted as a rejection of autonomy or lack of agency on the part of NHPI patients, but it could mean that autonomy and agency are exercised or expressed differently. For example, some key informants noted switching providers, registering complaints, seeking out alternatives to recommendations independently when they were not satisfied with their care.

Aspects of communication were also articulated more discreetly in the adapted version of the MOR index compared to the original. This reflects the insights from interviews in Phase 1 suggesting that, while visits typically allowed for explanations and questions, the NHPI participants in our sample appreciated more supportive and reciprocal communication. Elements highlighted included: feeling encouraged, feeling listened to, and feeling comfortable responding to questions asked.

The original MOR index also included three questions relating to verbal communication/conversation with providers (about level of comfort asking questions and accepting or rejecting recommendations), the culturally adapted version expanded that to include several more specific questions, some of which aligned closely with items included in the PCPC scale. Aspects that hindered open and caring communication were also articulated more specifically in the adapted version of the MOR index. Examples highlight patients’ perceptions of a risk of facing judgement, belittlement and feeling scared. These types of experiences and feelings are reflected in qualitive studies conducted with Native Hawaiian and Pacific Islander groups [18, 19, 21].

For this sample, the average adapted MOR index score was 132.5 (SD = 18.4) with scores ranging from 74 to 150. This suggests that, in general, participants reported a positive experience with their provider encounter during their maternity care. However, differences were observed by the characteristics of their prenatal care, with people who received less care having lower scores. These associations need to be interpreted with caution because of the small sample size as well as the lack of clarity about whether poorer evaluations of the patient-provider interaction would have generated fewer visits or fewer visits generated a poorer evaluation of patient-provider interactions. The same is true with first births, which could be related to differences in expectations of care between people experiencing birth for the first time compared to those who have prior experience. Yet, the topic warrants further investigation considering the importance of the first birth clinically and in terms of establishing a relationship between the healthcare system and the patient. Finally, the finding regarding the lower scores for community health centers (CHCs) could seem counter intuitive considering the important role CHCs have in providing primary care and prenatal care to low income, racialized communities, and other marginalized groups. Hawai‘i has 14 CHCs on six different islands, each with unique patient populations and programs. We did not collect detailed data on which CHCs patients attended, but we know that nationally CHCs face workforce shortages that can compromise their ability to provide the highest quality of care [41]. Hawai‘i has faced physician shortages for decades and experienced a substantial reduction in the physician workforce in 2020 [42]. Challenges accessing care, availability of particular services, staff burnout, and/or compassion fatigue could result from provider shortages at CHCs, possibly straining patient-clinician encounters and resulting in lower scores. Future studies should investigate this relationship in more depth.

The key strength of this study was the use of a multiphase process to culturally adapt and test the original MOR index and incorporated that included – content validity of items, cognitive interview testing, and factor analysis. However, this study did have certain limitations. First, the timeframe overlapped with the COVID-19 pandemic in Hawai‘i (the state’s first stay-at-home order went into effect on March 23, 2020). Changes in the healthcare system that were related to COVID-19, such as policies that limited the number of people who could attend the birth or increased burden on staff at the healthcare facilities may have influenced responses. It also could have generally raised the importance of having caring/emotionally supportive care for patients. However, we did not see differences in the overall score of the adapted MOR index between people who delivered during the height of COVID-19 (March 2020 – March 2021) in Hawai‘i and those who did not. Secondly, although Pacific Islanders represented half of the interview participants in Phase 1, the vast majority of survey respondents in Phase 2 were Native Hawaiian. This may have influenced the factor analysis if Native Hawaiians respond markedly differently to questions included in the 32-item MORi v2c. However, it is difficult to speculate how they might have influenced this phase of the research study. The size of the study may have limited the identification of specific cultural or social aspects among NHPI subgroups. Lastly, the interviews were conducted in English, which limited the participation of more recent Pacific Islander migrants who were not fluent in English. Work is ongoing to conduct a study within a much larger sample using the culturally adapted MOR index v3 (and translated versions) to better illuminate any differences between diverse NHPI subgroups.

Conclusion

Tools to measure NHPI patients’ experiences in healthcare are vital considering the stark disparities in outcomes and lack of clarity about how to reduce and eliminate inequities. The persistent finding that NHPI pregnant people participate less in prenatal care deserves deeper investigation into questions of perceptions of healthcare experiences and encounters with providers, factors that directly impact quality of care. The work conducted in this study provides a foundation for the next step in this research, which is to administer the culturally adapted survey in a larger sample of NHPI to determine if there are significant differences by subpopulation. This information can help inform future interventions which will be vital as NHPI populations continue to grow [43] and Pacific Islanders in particular, are increasingly settling in new, oftentimes rural communities, where their needs and values may not be clearly understood, compromising the quality of care they receive.

Abbreviations

NHPI

Native Hawaiians and Pacific Islanders

MOR

Mothers on Respect Index

PCPC

Person-Centered Prenatal Care

CBO

Community-based organizations

EFA

Exploratory factor analysis

CFA

Confirmatory factor analysis

DCC

Dignity, caring, and connection

PT

Perceived threats

PD

Perceived discrimination

Authors’ contributions

EL—methodology, analysis, writing & editing; AC – analysis, writing & editing; CVT – visualization, draft writing; JH, AD, SC, PLE – investigation, data acquisition, writing & editing; JK – conceptualization, funding acquisition, methodology, formal analysis, writing & editing; RD was involved in all aspects of the research project and manuscript development.

Funding

This work was supported by funding from the National Institutes of Health, including grants from the National Institute on Minority Health and Health Disparities (U54MD007601) and the National Institute of General Medical Sciences (U54GM138062). The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Availability of data and materials

Data that support the findings of this study were acquired through a community-engaged approach to research with Native Hawaiian and Pacific Islanders, indigenous peoples with a history of mistreatment in research. Datasets with over 40 unique records from this study may be made available from authors upon reasonable request and with permission of the community partners.

Declarations

Ethics approval and consent to participate

This study was approved by the Institutional Review Board of the University of Hawai ‘i (CHS# 2020–00627). Informed consent to participate was obtained from all participants in the study prior to data collection.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.