Adaptation and implementation outcomes of a parenting program for low-income, ethnically diverse families delivered virtually versus in-person

Abstract

Driven by the COVID-19 pandemic, many in-person health behavior interventions were compelled to quickly pivot to a virtual format with little time or capacity to reflect on or examine possible equity-related implications of a format that required digital access and remote learning skills. Using a parenting program for low-income families as a case study, this paper (a) outlines the process of adapting the program from an in-person to a virtual format and (b) examines the equity-related implications of this adaptation. Parents Connect for Healthy Living (PConnect) is a 10-session empowerment-focused parenting intervention designed to promote family health for Head Start families. In 2020, PConnect was adapted over a 6-month period from an in-person to a virtual format due to the advent of the COVID-19 pandemic. Three core elements were retained in the adaptation; session content, provision of coaching support for facilitators, and the co-facilitation model. Key modifications include session length, group composition, and language of program delivery. Head Start and PConnect records provided data to compare reach, acceptability, and appropriateness of virtual and in-person PConnect. Seventy-eight parents enrolled in the in-person program and 58 in the virtual program. Participant demographics and satisfaction were similar across formats, and demographics similar to the general Head Start population. Participation was higher in the virtual format. Parents participated in the virtual program primarily via smart phones (68%). This case study supports the acceptability and appropriateness of virtual parenting programs in ethnically diverse, low-resource settings.

Implications.

Practice: Virtual formats can be used to increase attendance rates in health promotion programs for low-income families.

Policy: Virtual health promotion programs must consider the frequency and timing of in-person sessions.

Research: Future research should be conducted to determine the relative effectiveness of a virtual or hybrid format, in comparison to an in-person format.

Introduction

Children and adults in lower income households are at heightened risk of chronic diseases such as diabetes [1], obesity [2, 3], and asthma [4, 5] compared to those in higher income households [1, 2]. The origins of such disparities are diverse and well-documented and include factors such as the neighborhood food and built environments (i.e., access to healthy foods and greenspaces) [6], environmental conditions (i.e., exposure to indoor and outdoor irritants) [7], and psychosocial distress (i.e., exposure to childhood adversity) [3, 8]. Because many chronic illnesses share common risk factors, low-income populations are also more likely to have multimorbidity [9]. Alongside strategies to address the underlying structural inequities at play, there is a critical need for effective health promotion programs that are tailored to the unique needs of low-income families [10, 11]. Such tailoring includes careful consideration of strategies to reach and engage low-income families, including implementation format.

With the advent of the COVID-19 pandemic, in-person health promotion programs were rapidly transitioned to, and subsequently delivered through, a virtual platform [12, 13]. Virtual intervention programs have demonstrated success in health promotion [14–16]; eliminating barriers to attendance (i.e., transportation and childcare) [17], while maintaining participant satisfaction [18] and effectiveness of the intervention [19, 20]. However, limited evidence exists on the acceptability and appropriateness of virtual programs for low-income parents of young children [13, 21]. It is critical to ensure that a virtual format does not create new or widen existing inequities in access to and use of such programs and unintentionally exacerbate existing health inequities. While a virtual platform has the potential to remove barriers low-income families face with in-person interventions, it is unclear if new barriers, such as low technology literacy and access, would arise. Although smartphone ownership remains high (>75%) for all Americans, socio-economic disparities persist in access to a computer and high speed internet at home [22]. Prior studies that examined the utility of a virtual intervention for low-income populations have shown promising results [12, 23]. However these virtual interventions were primarily asynchronous and therefore lack real-time interaction with group leaders and participants, which may be compromised when implemented in a virtual format.

Parents Connect for Healthy Living (PConnect) program is an empowerment-focused parenting program to promote healthy lifestyle behaviors and prevent chronic disease in low-income families. In response to the COVID-19 pandemic, PConnect was adapted in 2020 from an in-person to a virtual format. Using PConnect as a case study, we outline the process to adapt the group-based program to a virtual format and examine the equity-related implications of this adaptation. Guided by Proctor’s model of implementation outcomes [24], we examine differences in reach, acceptability, and appropriateness across the two formats. In doing so, we consider whether the potential benefits of a virtual format for low-income families outweigh the risk of introducing or exacerbating inequities.

Methods

Study design and setting

PConnect is 10-session health and empowerment program for caregivers of children enrolled in Head Start, a federally funded social service program created to break the cycle of poverty [25] by promoting school readiness for preschool-aged children from low-income families [26]. PConnect is part of an intervention, Community for Healthy Living (CHL), designed to prevent childhood obesity and promote child and family wellness in low-income families. Aligned with Head Start performance standards of fostering parent confidence and skills [27], PConnect has demonstrated effectiveness in increasing parent empowerment [28]. Informed by the family ecological model [29] and empowerment [30] theories, PConnect was co-designed by Head Start parents, staff, and the research team using community-based participatory research. PConnect is co-facilitated by a Head Start parent and staff member and covers a wide range of topics, including child health behaviors, family relationships, and parental advocacy. A summary of PConnect, the broader CHL intervention, and the randomized controlled trial to evaluate the intervention is provided in the protocol paper [31].

The CHL intervention, including the PConnect program, was evaluated in a cluster-randomized trial in 16 Head Start programs in the Greater Boston area serving over 1650 children and their families per year. Caregivers of children enrolled in Head Start intervention sites that were receiving the CHL intervention were eligible to participate in the in-person PConnect program in 2018 and 2019. Caregivers of children enrolled in control sites were not eligible to participate in in-person PConnect. Each session began with a community meal, which also served as an informal relationship-building and networking opportunity. For the remaining 90 minutes, the PConnect co-leaders reviewed the session topic(s) outlined in the facilitator’s manual using a combination of written materials (e.g., worksheets), dynamic group activities and group discussion. PConnect facilitators received coaching support throughout the program; after each session, the facilitators met with a member of the study team to debrief about the session, discuss any challenges that arose and plan for the next session. In spring 2020, the intervention trial was prematurely halted due to the advent of the COVID-19 pandemic. In a reactive adaptation [32], the research team adapted the in-person program over a 6-month period to a virtual format which took place over Zoom, a video web conferencing platform. Eligibility criteria for the virtual program mirrored the final year of the in-person intervention, where all 16 Head Start sites were included in the intervention. While, in-person, PConnect was conducted prior to the end of the intervention trial, virtual PConnect was implemented outside of the trial (fall 2020–spring 2021).

Adaptation process and rationale

In our adaptation, we sought to retain three intervention elements, including core intervention content, the provision of coaching support, and the co-facilitation model. Core intervention content included a learning component, a period of time where facilitators provided verbal and written information relating to the session topic (i.e., why is sleep important for children), and activities to apply this new knowledge. The session topics did not change and included nutrition and physical activity, sleep and screentime, child personality, healthy relationships, mindfulness, advocacy, and social networks. The learning component also remained unchanged, was discussion-based, and centered the parents’ experiences to empower them to make changes to improve family health. Decisions to modify or eliminate activities were made on a case-by-case basis and prioritized dynamic, interactive group activities, modifying only where necessary to ensure the appropriateness of a virtual setting. Curriculum developers reviewed each activity, taking into consideration how each activity would translate to a computer/smartphone screen, the length of time the activity took, and the social/emotional needs of parents. The coaching model was retained. Finally, we were able to retain the co-facilitation model in half of the groups. The remaining groups were facilitated by Head Start staff members and, in one instance, members of the research team, as it was difficult to recruit parent facilitators for the virtual format in a language other than English. Table 1 summarizes intervention elements that were modified and the rationale behind the changes. The most noteworthy change was the reduction of each 2-hr session (30 min meal + 90 min content) to 1 hr, in recognition of “Zoom fatigue.” While there was not a body of research on demographic differences in Zoom fatigue to draw on at the time, the research team was sensitive to the possibility that Zoom fatigue may be disproportionately experienced by parents of young children from low-income backgrounds, given lower technology access [22] and digital literacy reported in low-income populations [13, 33]. Other modifications were relatively minor.

Table 1.

Summary of changes as the in-person program was adapted for the virtual format

Modifications to accommodate the virtual format		Rationale
Session length	Was reduced from 2 hr to 1 hr.	To prevent Zoom fatigue.
Group composition	Expanded from only including parents from a single Head Start center to including parents from any participating centers in Greater Boston.	To increase program access. This allowed parents to choose a program to attend based on the time, day of the week and language.
Time of day sessions were held	In-person sessions were held on weekday mornings (when preschool-aged child was attending Head Start). Virtual sessions were conducted at a range of times throughout the day.	To increase program access. Parents could choose from a selection of session times to meet their work and childcare needs.
Weekly frequency of sessions	Was modified from once per week to either once a week or twice a week per parent preference.	To increase convenience. Virtual format allowed for flexibility in the frequency of sessions as there was no travel time required of parents or staff.
Number of languages utilized within a group	Some in-person groups included parents who spoke different languages (e.g., monolingual in English or Spanish). In these sessions, a multilingual facilitator or parent translated the content. All virtual sessions were offered in English or Spanish.	To simplify sessions and ensure sessions were not disjointed by having to repeat content in another language.
Community meal	Instead of the community meal utilized in the in-person format, parents received a $10 grocery gift card for each session they attended.	Equity in program benefits. To alleviate financial stress experienced and provide compensation their time and effort, as a community meal is not feasible virtually.
Technical support	Technical support was provided (e.g., managed the PowerPoint slides and Zoom chat, technology trouble shooting) by a research assistant during the virtual program; no technical support was needed in-person.	To eliminate technology issues as a barrier for participants accessing the program and for facilitators delivering session content.
Hard copy versus online materials	For in-person sessions, only hard copies of session materials were provided. For the virtual program, a study website was created with all program materials/resources. Materials were available online and hard copies could be requested if desired.	To facilitate access to program materials.
Tablets were offered	Tablets were offered to all participants in the virtual group that could be used to access PConnect sessions. Flexibility was also provided in the device use join PConnect, allowing parents to join with their smartphone from varying locations.	To eliminate barriers to access. Of note, no virtual program participant requested to borrow a tablet.

Data collection

Data for this study were compiled from a range of sources. The data were collected for the in-person program across two academic years (2018–2019, 2019–2020) and one academic year for the virtual program (2020–2021). Similar data were compiled for the program formats. Demographic data (gender, race/ethnicity, primary language, employment status, and number of parents in the household) were extracted from Head Start enrollment records for all PConnect participants and for all Head Start parents who had a child enrolled in one of the participating Head Start programs during year 2 of the trial. PConnect attendance was tracked across both formats. Parents who attended sessions 9 or 10 were invited to complete an anonymous exit survey; survey questions, available in English or Spanish, solicited feedback on participant satisfaction with the program. Fifty-three percent (41/78) of parents from the in-person and 74% (43/58) from the virtual program completed the exit survey. Parents who graduated from PConnect were invited to participate in a semi-structured interview that was administered in English, Spanish, or Chinese by a trained research assistant. Interview questions addressed relationships formed with other parents and satisfaction with the program. Twenty-one percent (16/78) of parents from the in-person and 36% (21/58) from the virtual program participated in the semi-structured interview. In spring 2021, all parents with a child enrolled in Head Start were invited to complete an online survey assessing access to technology. Forty-one out of 58 (71%) parents participating in virtual PConnect completed this survey along with 55 parents not participating in PConnect.

Informed by Proctor’s Conceptual Model of Implementation Outcomes [24], the data compiled were used to operationalize three implementation outcomes: (1) reach, (2) acceptability, and (3) appropriateness. The definitions for each implementation outcome and the data sources utilized are outlined in Table 2.

Table 2.

Definition and operationalization of implementation outcomes

Implementation outcome	Definition	Data sources	Variables
Reach	The percentage and characteristics of individuals who participate in the intervention in comparison to the population served (i.e., all individuals who are eligible to participate) [47].	PConnect attendance sheets Head Start enrollment records PConnect attendance sheets Head Start enrollment records	Number of parents who enrolled in the in-person and virtual PConnect programs and proportion of eligible parents Demographic characteristics (race/ethnicity, primary language, employment status, and number of parents in the household) of in-person and virtual PConnect participants compared with eligible parents
Acceptability	Stakeholder satisfaction with the content and complexity of the intervention [24].	PConnect attendance sheets PConnect exit surveys and semi-structured interviews	Number of sessions attended and graduation rates for in-person and virtual PConnect participants Participant satisfaction with in-person and virtual PConnect programs
Appropriateness	Stakeholder’s perceived fit and relevance of the intervention [24].	Head Start parent survey PConnect exit surveys PConnect exit surveys and semi-structured interviews PConnect semi-structured interviews	Access to technology for virtual PConnect participants and Head Start parents in general Ease of technology use, use of technology, virtual PConnect only Relationships formed among in-person and virtual PConnect participants Facilitators/barriers to in-person and virtual PConnect program attendance

Data analysis

We used a convergent, mixed-methods approach to examine differences in implementation outcomes across the two formats; the quantitative and qualitative data were collected and analyzed separately and, after individual analysis, were integrated in narrative form [34] using a side-by-side comparison [35, 36]. For program reach, we compared the demographic characteristics of Head Start parents, in-person PConnect participants, and virtual PConnect participants. Acceptability was evaluated by comparing attendance and satisfaction between the two program formats and examining attendance rates by race/ethnicity. Appropriateness of the virtual platform was assessed by examining technology access among parents who participated in virtual PConnect compared with Head Start parents, in general, and examining relationships formed between facilitators and barriers to attendance for in-person and virtual PConnect parents.

All qualitative interviews were audio recorded, transcribed verbatim, and translated into English (for interviews conducted in Spanish or Chinese). Two coders (NG and AA) coded the transcripts in NVivo version 11 using an inductive–deductive thematic analysis approach [37]. First, a codebook was created based on our research questions. Second, the two coders independently coded five transcripts, and then met to examine discrepancies. Discrepancies were discussed until an agreement was reached on the codes’ definition and interpretation. Next, the two coders independently coded nine transcripts, convened, and examined inter-rater reliability. Due to an acceptable inter-rater reliability (kappa coefficient 95.72), one coder (NG) coded the remaining 23 transcripts. The second coder (AA) then reviewed these 23 transcripts and corresponding codes to check for accuracy. Finally, a detailed reading of each code was conducted to summarize codes, develop themes, and identify illustrative quotes.

All procedures were approved by the Institutional Review Board.

Results

As summarized in Table 3, 136 parents enrolled in PConnect and attended at least one content session (sessions 2–9) between 2018 and 2021; 78 parents in the in-person and 58 in the virtual program.

Table 3.

Summary implementation characteristics of in-person and virtual PConnect programs

	In-person PConnect	Virtual PConnect
	n (%)	n (%)
Total parents who attended at least one session	78	58
Number of PConnect groups held	12	9
Number of PConnect enrollees per group
2–4 participants	5 (42%)	3 (33%)
5–8 participants	3 (35%)	4 (44%)
9–12 participants	3 (35%)	2 (22%)
13–16 participants	1 (8%)	0
Average number of parents per group	6.5	6.1
Implementation language(# of groups)
English	6	5
Spanish	1	4
Chinese	1	0
English-Spanish bilingual	4	0
Frequency of sessions
Weekly	12	3
Twice a week	0	6
Time of day offered
Morning	12	3
Lunch time	0	2
Afternoon	0	2
Evening	0	2

Reach

Overall, the demographic characteristics of PConnect participants reflected the demographics of the population of Head Start from which they were drawn (Table 4). While the proportion of eligible families who enrolled in PConnect was low (approximately 5% in-person, 4% virtual), there was no evidence that the PConnect program selectively reached certain subgroups of Head Start families. While the parents who participated in each program format were very similar, there were a number of differences. For instance, parents who attended in-person PConnect were predominately Spanish-speaking (55%), whereas, the majority of participants in the virtual program had a primary language other than English or Spanish (48%), in most instances, Haitian Creole (12%). In addition, slightly more parents in the in-person program were from two versus one parent households (58% from two parent; 38% from one parent); this was not true for parents in the virtual program for which the percentage from one versus two parent households differed only slightly (52% vs. 48%, respectively). Lastly, a somewhat higher percentage of parents in the virtual program (38% virtual, 29% in-person) were unemployed.

Table 4.

Demographic characteristics of Head Start parents and parents who participated in the PConnect program, in-person, and virtual formats (reach)

	Head Start Year 2a n (%)	In-person PConnectb n (%)	Virtual PConnectc n (%)
	n = 1,825	n = 78	n= 58
Gender
Male	94 (5%)	3 (4%)	5 (9%)
Female	1720 (94%)	74 (95%)	52(90%)
Missing	11	1	1
Race/ethnicity
NHd White	140 (8%)	4 (5%)	6 (10%)
NH Black	642 (35%)	14 (18%)	14 (24%)
Hispanic/Latino	800 (44%)	46 (59%)	23 (40%)
NH Asian	165 (9%)	10 (13%)	12 (21%)
Other	42 (2%)	2 (3%)	1(2%)
Missing	36	2	2
Primary language
English	585 (32%)	14 (18%)	8 (14%)
Spanish	660 (36%)	43 (55%)	21 (36%)
Other	561 (31%)	20 (26%)	28 (48%)
Missing	19	1	1
Employment
Unemployed	490 (27%)	23 (29%)	22 (38%)
Other	1323 (72%)	52 (67%)	33 (57%)
Missing	12	3	3
Number of parents in the household
1	1247 (68%)	30 (38%)	30 (52%)
2	578 (32%)	45 (58%)	28 (48%)
Missing	25	3	0

Percentages may not add up to 100 due to missing data.

data from Head Start parents who had a child enrolled in one of the 16 participating Head Start centers during year 2 (trial midpoint) of the trial,

data from year 1 (2017–2018) or year 2 (2018–2019),

data from year 4 (2020–2021),

non-Hispanic.

Acceptability

The average number of sessions participants attended was higher in the virtual program (7.1) compared to the in-person program (5.8) (Fig. 1A). Stronger racial/ethnic patterning in program acceptability was observed in the in-person compared with the virtual format. For the in-person format, non-Hispanic Asian parents had the highest average attendance rates (8.0). Hispanic/Latino, non-Hispanic Black, and non-Hispanic White parents had notably lower attendance rates attending 5.9, 4.8, and 3.8 sessions, respectively. The average number of sessions parents attended for the virtual program was comparable for non-Hispanic White (7.8), non-Hispanic Black (7.6), Hispanic/Latino (6.6), and non-Hispanic Asian (7.1) parents. Similar results were found for graduation rates (Fig. 1B).

Fig 1.

(A) Average number of PConnect sessions attended, in-person, and virtual formats. (B) Percentage of parents who graduated from PConnect (attended ≥ 70% of session), In-person and virtual formats.

Satisfaction

In terms of satisfaction, across both formats, 90% or more of parents reported that PConnect was useful and that they would recommend the program to others. Findings, across both formats, from the qualitative interviews confirmed high levels of program satisfaction and engagement with participants stating content was informative and relatable (Table 5). Findings converged across quantitative and qualitative data (Table 5); parents reported high levels of satisfaction with PConnect with few differences observed across formats.

Table 5.

Joint display used to compare and integrate findings from qualitative interviews and quantitative data

Qualitative themes	Illustrative quotes	Quantitative data
			In-person PConnect n (%)	Virtual PConnect n (%)
Satisfaction	“It was a productive experience. I was productive, encouraging, and I think you should do it more often. It is a very interesting program. The information is super good and it’s important and productive. I was fascinated.” [In-person PConnect, Hispanic participant] “Each session is different, so there’s always something you look forward to. You look forward to learning. I think that’s what kept me interested. ‘Okay, that one was great. The next one will be about another subject. So, I’m sure it’s gonna be better, so I have to be there to learn.’ They’re all unique. That’s what I would say kept me coming back.” [Virtual PConnect, non-Hispanic Black participant]	PConnect Useful Recommend PConnect to Others	41 (100%) 41 (100%)	39 (97%) 38 (100%)
Barriers/facilitators	“Because sometimes it’s not easy to go to whatever […]. Sometimes, there’ll be no time to even go there. When it is online, you could be anywhere. Wherever you are. If you are at work, you could just logon when you’re free.” [virtual PConnect, non-Hispanic Black participant] “I felt that online was easier because I didn’t have to worry about who can I leave my girls with or who will go pick them up from school.” [virtual PConnect, Hispanic/Latino participant]	Easy to Use Zoom Easier to Participate Virtually than in Person Felt Engaged in Sessions Disruptions at Home Limited Engagement		36 (95%) 32 (84%) 31 (82%) 10 (26%)
Relationships	“Yes, I get on very well with one of them. Actually, yesterday I picked up her daughter because she works very early. I said, ‘Don’t worry I can pick her up.’ We always talk on WhatsApp. I know two of them, but I’m closer to one of them and talk to her more often.” [in-person PConnect, Hispanic/Latino participant] “It was interesting. Somehow, I felt that there was more freedom. Maybe in person they wouldn’t have dared to say, ‘My child wasn’t that good, I went through this, and my daughter is...’ In person we might have been ashamed or wouldn’t have dared. Online, however, we had that freedom to express ourselves without feeling shame or embarrassment.” [virtual PConnect, Hispanic/Latino participant] “Good thing about joining the program online. The good thing is that you are in your comfort zone. You’re comfortable where you are because I may be with them if it were to be in-person. I’m a shy person. I may not even participate (if it was in person). Because I’m here. I know we are not really together. I could talk. I could say whatever I wanna say. I think it was easier online for me than in-person.” [virtual PConnect, non-Hispanic Black participant]	Built Relationships over Zoom FeltOther Parents Participated in PConnect Believed Parents Supportive of One Another Planned to Stay in Touch with One Another PConnect Sensitive to Values/Culture	39 (93%) 31 (78%) 29 (73%)	31 (82%) 31 (82%) 38 (100%) 27 (71%) 30 (81%)

Missing data were excluded from analysis,

All data reported from the parent exit survey.

Appropriateness

Eighty five percent of Head Start parents (inclusive of Head Start parents who participated in PConnect and those who did not) who completed the survey reported they had access to a smart phone, 73% had access to the internet at home, and 66% had access to a tablet or laptop (Table 6). Compared to Head Start parents overall, those who participated in the virtual PConnect program had comparatively lower access to technology with 73% reporting access to a smart phone, 61% the internet, and 41% a tablet or laptop. Parents who participated in the virtual program also reported the device they used to participate (data not shown). Approximately two-thirds of parents used their smart phones to participate in the program.

Table 6.

Access to technology in the home among Head Start parents and virtual PConnect participants (appropriateness)

	All	NH White	NH Black	Hispanic/Latino	NH Asian
	n (%)	n (%)	n (%)	n (%)	n (%)
Head Start parents from participating programs
Number completing survey	96	14	26	31	23
Access to technology at home
Access to smartphone	82(85%)	12 (86%)	23 (88%)	26 (84%)	19 (83%)
Access to tablet or laptop	63(66%)	12 (86%)	17 (65%)	14 (45%)	17 (74%)
Access to internet	70(73%)	11 (79%)	19 (73%)	19 (61%)	17 (74%)
Virtual PConnect
Number completing survey	41	5	8	18	9
Access to technology at home
Access to smartphone	30(73%)	4(80%)	6(75%)	13(72%)	6(67%)
Access to tablet or laptop	17(41%)	3(60%)	4(50%)	3(17%)	6(67%)
Access to internet	25(61%)	5(100%)	6 (75%)	8(44%)	5(56%)

Missing data were excluded from analysis. Three individuals (Head Start n = 2; virtual n = 1) identified as “Other” racial/ethnic group and were not included in analysis that were broken down by race/ethnicity. All data reported are from the parent survey in spring of 2021.

Barriers/Facilitators

Parents who participated in the virtual program provided feedback on the ease of the virtual format (Table 5). The vast majority (>80%) of parents reported that it was easy to use Zoom, it was easier to participate virtually than in person, and they felt engaged in the sessions (Table 5).

Qualitative data from the semi-structured interviews provide additional insight into the barriers and facilitators to attendance. Parents reported that barriers to attending in-person sessions were conflict with work schedule and lack of childcare for younger children or transportation. The vast majority of the parents from the virtual program stated that it was easier to participate online than if PConnect were held in person. Few mentioned technology difficulties were a barrier to participation, and several cited benefits to participating in a virtual program, such as the ability to balance parenting responsibilities with Zoom classes. Unique to a virtual format is the ability to make up missed sessions. One parent we interviewed was able to attend another PConnect group, because the virtual sessions were not directly tied to a Head Start site, and two parents recommended recording the sessions so parents who were absent had access to the content. Findings from qualitative and quantitative data from participants in the virtual program converged supporting the advantages of attending a virtual program.

Relationships

The data compiled suggest that parents across both formats had the opportunity to build relationships and to strengthen their social networks (Table 5). Specifically, most (>90%) participants from the in-person and virtual programs reported that parents were supportive of one another with the majority (>70%) stating they planned to stay in touch with other parents in PConnect. Over 80% of parents from the virtual format stated that they were able to build relationships over zoom and also felt other parents participated in PConnect (Table 5).

Qualitative findings from the semi-structured interviews reveal that participants, across both formats, learned from, shared resources with and felt supported by one another. Parents who participated in the in-person program built relationships that extended beyond the individual PConnect sessions. Outside of PConnect, these parents saw one another in-person and communicated over a messaging app to share resources and provide emotional, social, tangible, and parenting support. These interactions were both planned and unintended, as many parents saw one another at Head Start or in the community. While parents in the virtual program did not build strong friendships and rarely communicated outside of PConnect, the virtual program allowed participants to feel confident and enabled them to communicate and connect with other participants in an authentic way. These participants believed that the relationships formed and support received from other participants would not have been as strong if the program was held in person.

Findings from both qualitative and quantitative data support the formation of strong relationships in the in-person format. Findings from the two data sources on relationships diverged slightly for participants in the virtual format, with qualitative findings illustrating the nuances of virtual relationships and how relationships and parental support differ across the two formats. In contrast to the in-person format, parental support and relationships in the virtual program were confined to the PConnect sessions.

DISCUSSION

Overall, findings indicate that it is appropriate and acceptable to implement a virtual parenting program for Head Start parents or similar programs serving parents in low-resource settings. The majority of findings converged across quantitative and qualitative data supporting high levels of satisfaction with PConnect and advantages of accessing a virtual format. Qualitative data on relationships expanded findings from quantitative data with qualitative data providing additional

insight into how parental support and relationships differ across the two formats. Importantly, moving to a virtual context did not appear to introduce inequities in implementation. In fact, many changes associated with the virtual format (e.g., group composition, session time of day, session frequency) served to increase access to the program.

In terms of reach, the demographic characteristics of parents in the virtual and in-person programs were similar and also representative of the Greater Boston Head Start population; that is, there was no evidence that the PConnect program selectively reached certain subgroups of Head Start parents. This is consistent with previous reports of minimal demographic differences when comparing in-person to virtual implementation [19, 38–40]. The exception in our study is that most in-person participants identified as Hispanic/Latino and primarily spoke Spanish. There was slightly more diversity among parents in the virtual format, in which most participants, had a primary language other than English or Spanish (e.g., Haitian Creole). Greater ethnic homogeneity in the in-person format was associated with eligibility by sites located in different neighborhoods, which was not a factor in the virtual format.

Although rates of satisfaction and engagement were comparable across both formats, attendance, used as a measure of acceptability, was higher among participants of the virtual format. This is likely because the modifications made to PConnect to accommodate the virtual format naturally allowed for more flexibility. For instance, a 1-hr reduction in session length and offering sessions during lunch and evening hours in and of itself made the program easier to access for parents who have competing responsibilities. This was true for all racial/ethnic groups except non-Hispanic Asian parents. To accommodate parents’ linguistic ability at one Head Start center, one in-person group was conducted in Chinese which likely drove higher rates of participation for the in-person format. A Chinese language session was not conducted virtually because parents were not clustered into groups based off their geographic location. While prior findings on satisfaction and attendance are not specific to low-income populations, our findings align with prior literature examining satisfaction [19, 39, 41] and add to the mixed results on findings regarding attendance; some concluding there are no differences in attendance between the two formats [38, 42] and one study finding higher rates in the virtual format [40].

Our initial concerns that virtual programming might increase inequities due to lack of either internet access or remote learning skills were not supported [22]. Since Head Start, in general, including PConnect, serves low-income families a less accessible virtual format would presumably exacerbate social disadvantage. This mirrors prior studies, although not focused on low-income families, on tele-health that have confirmed a tele-health platform is cost-effective, convenient, and accessible for parents [41, 42]. In addition, remote learning and working were part of the overall response to the pandemic, which may have facilitated the effectiveness of the virtual format which was implemented after remote education and efforts to promote digital equity (i.e., providing families with internet, Chromebooks) were widespread in the area [43].

Relationships are central to the underlying goals of PConnect, as participant interaction provides an opportunity to increase engagement and group cohesion, both of which are important for supporting behavior change [44]. Although the types of relationships formed and support received in the virtual and in-person programs differed, parents in the virtual program felt supported by and successfully shared resources and information with one another. This is a noteworthy finding, as few studies have examined the differences in the quality of relationships formed across an in-person and virtual format.

These findings are not without limitations. First, the in-person and virtual formats were implemented in very different contexts; prior to and during the COVID-19 pandemic. We recognize that the context alone might explain the differences in implementation outcomes, particularly acceptability. Given that COVID-19 disproportionally affected low-income workers [45], there may have been more motivation, and in turn higher attendance rates, to engage in community programs. Regardless of this possibility, as we move forward into an era where remote work, education, and relationships are a social norm, we must understand how a virtual era benefits or hinders the implementation of interventions in low-resource settings. A second limitation is the small sample size. While many studies to date comparing in-person and virtual implementation of programs [19, 41, 42] have had small sample sizes, it nonetheless limits our ability to draw conclusions. In addition, we examined the outcomes by race/ethnicity in order to determine if a virtual format introduces inequities. The small sample size precludes our ability to analytically test differences between the racial/ethnic groups. Third, satisfaction with PConnect, relationships formed, and ease of technology use were measured during sessions 9 and 10, thereby limiting its distribution to only who graduated from PConnect. This is problematic if participants dropped out of PConnect due to dissatisfaction with the program and/or technology issues, as their responses were not captured. Additionally, the parent survey to assess access to technology was distributed in electronic form. This modality of delivery alone introduces bias and possibly overestimates Head Start parents’ access to technology. With some Head Start centers being fully remote and others adopting a hybrid model in the spring of 2021, paper surveys were not a feasible option.

Conclusion and Future Directions

The overarching goal of this work was to inform future discussions of the benefits and limitations of virtual programs for low-income families. Based on PConnect, our findings suggest that a virtual platform is an acceptable and appropriate mode of delivery for health promotion programs that target parents in low-resource settings with few differences in implementation outcomes observed by race/ethnicity. To meet the diverse needs of low-income parents, future parent empowerment programs should consider a hybrid approach. This approach will couple the flexibility of a virtual format with the ability to build strong relationships that allow communication and support to extend outside the boundaries of the program. To further increase flexibility and access, all virtual session content should be formatted to and accessible with a smartphone. In addition, programs should offer online materials and consider recording virtual sessions and/or allowing participants to attend a different group’s session so participants who are unable to attend a session can access the content.

While a hybrid approach has the potential to further increase engagement and participant success [46], the frequency and timing of which in-person meetings should be conducted is unclear. PConnect graduation, which occurs during the last session of the program, is a natural point for an in-person session to occur. However, in order to build group cohesion and foster participant relationships that extend outside of the bounds of PConnect, in-person meetings may also have to occur at an earlier stage of the group. Future research must be conducted, in conjunction with parents, to determine the most effective and accessible structure of a hybrid model for parents in low-resource settings. Finally, as our results only shed light into the acceptability and appropriateness of a virtual parent program, future research on the comparative effectiveness of virtual versus in-persons programs in low-resource settings is needed.

Funding

This work was supported by the National Institutes of Health (grant number R01DK108200).

Compliance with Ethical Standards

Conflicts of Interest: All authors declare no conflicts of interest.

Human Rights: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed Consent: Informed consent was obtained from all individual participants included in the study.

Welfare of Animals: This article does not contain any studies with animals performed by any of the authors.

Transparency Statements: Communities for Healthy Living was registered at ClinicalTrials.gov, NCT03334669 in October 2017. The analysis plan was not formally pre-registered. De-identified data from this study are not available in a public archive. Analytic code used to conduct the analyses presented in this study is not available in a public archive. They may be available by emailing the corresponding author. Materials used to conduct the study are not publicly available.