Home-visiting programs are an often-used intervention for supporting the well-being of families with young children, and are typically targeted to families experiencing risks and adversities (Heaman et al., 2006; Korfmacher et al., 2008; McKelvey et al., 2016). Home-visiting programs are used for a variety of purposes, such as supporting new mothers with infants, promoting early parenting skills, improving economic self-sufficiency, reducing child abuse, and teaching children early readiness skills, and are strongly endorsed by the American Academy of Pediatrics as an effective practice for enhancing children’s well-being (Duffee et al., 2017).

Given the ongoing expansion of home-visiting programs in a number of locales, coupled with the non-trivial costs of such programs, there is long-standing interest among researchers in ensuring the effectiveness of this practice. Researchers have conducted a number of meta-analyses of home-visiting programs to evaluate estimates of impacts on such outcomes as parenting behaviors, child cognitive outcomes, child physical development, and child maltreatment, among others (Filene et al., 2013; Kendrick et al., 2000; Nievar et al., 2010; Sweet & Appelbaum, 2004). Although such reviews find generally positive (albeit small) impacts of home-visiting programs on key outcomes (e.g., Filene et al., 2013), in general these studies find that program effects are inconsistent across the individual studies reviewed. In part, this lack of consistent effects may reflect difficulties in ensuring that families receive the desired “dosage” of home-visiting activities, which is considered a critical component of effective home-visiting programs (Nievar et al., 2010). In part, failure to provide home-visiting activities at the desired dosage occurs because such programs often target families experiencing disadvantage and thus may be difficult to retain in program activities. As is well documented, familial disadvantage is associated with a range of stressors that can impede families from maintaining engagement in home-visiting activities, including those associated with economic insecurity (e.g., working unconventional hours, searching for employment, frequent moves) (Rothwell & Han, 2010; Wadsworth, 2012).

Supporting treatment completion throughout implementation substantially enhances potential intervention effectiveness (Sanetti et al., 2021). Studies on treatment completion find that children who prematurely dropped out of treatment presented with considerably more symptoms to be treated, and their caregivers reported experiencing significantly more parenting stress and depressive symptoms (Pellerin et al., 2010). Moreover, likelihood of treatment completion is linked to caregivers’ demographic characteristics (Fagermoen et al., 2023; Sieger et al., 2023), with completers tending to be female, younger, college educated, and white. Such socioeconomic disparities in treatment completion are similarly reflected in preschool enrollment rates and kindergarten readiness skills (Crosnoe et al., 2016). The Family Stress Model is an explanatory framework, proposing that conditions of poverty cause parental stress which disrupts parenting quality, leading to poorer social and academic child outcomes (Justice et al., 2019; Masarik & Conger, 2017). Families and children experiencing an accumulation of disadvantages are therefore simultaneously the most likely to be targeted for home-visiting programs, but least likely to be retained.

The nationwide Maternal, Infant, and Early Childhood Home-Visiting (MIECHV) program began in 2010 to provide funding for the development, implementation, and evaluation of evidence-based home visiting programs (U.S. Department of Health and Human Services, 2019). Since then, Home-Visiting Evidence of Effectiveness (HomVEE) systematic reviews are published annually, estimating program impacts based on specified criteria. High rates of attrition in home-visiting studies lends itself to lower impact ratings. With the national attrition rate of MIECHV participants being over 50% (Duggan et al., 2018), it is critical that factors associated with retention or attrition of participants are identified. Individual studies identify associations between family characteristics and retention in MIECHV programs, but findings remain inconsistent across studies (Bower et al., 2020).

Researchers have rigorously investigated retention-specific difficulties within home-visiting programs with respect to the Nurse-Family Partnership, a MIECHV home-visiting program that targets first-time mothers in low socioeconomic status (SES) homes (Olds, 2006). The Nurse Family Partnership is a scaled approach designed to support first-time mothers via home visiting conducted by nurses in community-based settings. A number of studies have explored retention and dropout in the Nurse Family Partnership programs to identify factors that promote retention in terms of the providers and families served (Beam et al., 2010; Holland et al., 2014a, 2021). In turn, the presence of kindergarten-transition difficulties at school entry can impede learning over the kindergarten academic year, with different children facing unique and individualized challenges.

Home-visiting programs offer a means to prepare parents and children for kindergarten entrance and potentially ease this transition, with a particular focus on children experiencing risk factors that compromise school readiness, such as poverty. Bierman and her colleagues, for instance, implemented home visiting with caregivers of children attending Head Start, a federally supported program for 3- to 5-year-old children residing in low-SES homes (Bierman et al., 2008). In this randomized controlled trial, Head Start teachers implemented a variety of evidence-based strategies for promoting children’s academic and social-emotional skills while children’s caregivers received ongoing home visits targeting the same set of skills. In total, caregivers received up to 16 home visits in the spring of preschool (10 sessions) and through the forthcoming fall, when children were in kindergarten (six sessions) (Bierman et al., 2021). Longitudinal examination of program impacts indicates that the classroom practices plus home visiting resulted in long-term positive impacts on children’s reading and social skills while also reducing children’s future needs for special services (Bierman et al., 2018). These sustained effects may reflect the increased home-learning activities reported by caregivers receiving home visits (Bierman et al., 2015) or caregivers’ heightened expectations towards their children’s academic development (Loughlin-Presnal & Bierman, 2017).

In a more recent study, Purtell et al. studied impacts of the Kindergarten Transition Program (KTP), a comprehensive intervention program adapted from a published manual (Pianta & Kraft-Sayre, 2003) focused explicitly on improving children’s kindergarten transition (Purtell et al., 2022). Instead of explicitly seeking to build children’s readiness skills, KTP focuses on ecologically informed practices to build connections between families and schools in an effort to create strong, long-lasting influences on children and parents’ schooling experiences. In this work, KTP was introduced into a subset of classrooms in two urban communities’ expansion of preschool programs targeting very low-income families.

Based on a randomized controlled trial research design, 52 preschool classrooms serving preschool-aged children from low-SES homes were assigned to implement KTP or maintain business-as-usual practices in the first cohort of the study. KTP implementation involved classroom teachers and transition coordinators delivering a range of transition-supporting practices, such as visiting kindergarten classrooms, holding kindergarten informational sessions for parents, and implementing structured lessons focus on kindergarten routines and rituals. Further, children within KTP-assigned classrooms were randomized to receive a home-visiting program (referred to as KTP +) in which transition coordinators worked directly with parents from the fall of preschool to fall of kindergarten to prepare for the kindergarten transition and then, after the transition took place, support parents in building connections with the kindergarten milieu. In total, parents in the KTP + condition were to receive nine home visits over a 15-month period. An initial investigation of KTP impacts on parent outcomes showed that the home-visiting program (KTP +) had significant, positive impacts on parents’ school engagement for Black and Hispanic parents, but not white parents (Purtell et al., 2022).

Though prior studies of home-visiting programs designed to promote school readiness report general attrition information, they typically do not detail the extent to which parent participation was sustained over time, nor the characteristics of families who were successfully retained. Prior research on home-visiting programs targeted to very young children and their parents, such as the Nurse Family Partnership, consistently indicate that retention of parents is challenging and that certain subgroups of parents are most vulnerable to attrition (Beam et al., 2010; Holland et al., 2014a, 2014b; O’Brien et al., 2012; Olds et al., 2013). However, the literature focused on subgroup attrition in home-visiting programs to promote school readiness among preschool-aged children is less clear. In the Bierman et al. (2008) study, authors reported that parents completed an average of 12 of 16 sessions overall (75%, on average); in the Purtell et al. (2022) study, families received 3.4 visits on average (range of 0 to 5; 70% on average), and that only 54% of the sample completed all five home visits, but neither study reports the characteristics of families that attrite versus complete the program. Such findings suggest the need to better understand patterns of participation and loss among families in such programs, as well as family characteristics that may relate to participation outcomes.

Given that the accumulated evidence indicates that attrition is a significant risk in home-visiting programs (Azzi-Lessing, 2011), it is necessary to understand the extent to which parents involved in home-visiting programs targeting preschool-to-kindergarten transition can be retained and supported over time. Importantly, should the positive impacts of home-visiting programs during the preschool-to-kindergarten transition only reflect those who are maintained in the program over time, impact estimates are likely biased. Thus, the present study used data from the KTP project to examine caregiver retention during a home-visiting program transcending their children’s preschool year into the fall of kindergarten. During this 15-month home-visiting period, caregivers were to receive an intended nine home visits from transition coordinators assigned to support KTP implementation among participating families. As the focal variable of the current study, the number of completed home visits and the timing of the last home visit were recorded for each family. A family is classified as a dropout if they did not complete any home visit, or if their last completed home visit occurred before the ninth visit. In other words, families who dropped out stopped participating before the final home visit.

Study aims were threefold. First, we sought to examine home-visit participation levels among families and explore correlates of participation levels as measured by the number of completed home visits. Second, we sought to determine the number of families who were retained in the home-visiting program and to determine the extent to which these families differed from those who attritted using group-based comparison. Finally, the third aim was to evaluate patterns of attrition over time using survival analysis, and to investigate child and family characteristics that potentially contribute to survival, or completing the program. In particular, in all three aims, we examined a range of child-family demographic characteristics (gender, age, race, ethnicity, home language, maternal highest level of education attained, household income, IEP status, size of the household, and number of caregivers in the household) and children’s academic skills in the fall of preschool. These attributes are commonly examined as predictors or covariates in literature on family-based intervention and some variables have been shown to correlate with treatment efficacy or program participation (e.g., Bower et al., 2020; Miller et al., 2008; LoBraico et al., 2021). Therefore, the current study can be viewed as exploratory in identifying potential predictors of retention in home visiting programs.

Method

Research Design and Overview

The present study uses data from the first of two intended cohorts of the Kindergarten Transition Program (KTP; Purtell et al., 2022), a randomized controlled trial that aims to support preschool-aged children’s transition into formal schooling and spanned the transition from pre-kindergarten to the fall of kindergarten. KTP incorporates multiple components of treatment, with home visiting program being one of the key intervention components for families assigned to the KTP + intervention condition. The first cohort participated from fall of 2018 to late fall of 2019, and the second cohort participated from fall of 2019 to late fall of 2020. Unfortunately, the second cohort experienced disrupted research activities and eventual termination due to the global pandemic. Consequently, the analyses presented in this study only involve the first full cohort to address the three study aims, and the study represents participation patterns among families during pre-pandemic conditions. As discussed above, the focus of the main study was to investigate the impact of KTP on family connections and child outcomes based on the full sample, whereas the current study focuses on the patterns and correlates of participation in the home-visiting program using a subsample of families who were randomly assigned to receive the intervention.

Participants

Classrooms

This study involved preschool classrooms located in two large urban areas that were actively expanding preschool enrollment through various initiatives, with a specific focus on enrolling children from under-served groups, including families living in poverty. We used the term “preschool” instead of “pre-kindergarten” to describe the classrooms, as they primarily served children between the ages of 3 and 5 years old.

The first cohort of this study comprised 52 classrooms in 29 different preschool programs, including two large urban public-school districts, a tri-county Head Start network, and a private subsidized child-care center, all located in Ohio. About one-half (51%) of the programs operated on a full-day, full-week schedule; 31% were full-day, four-day week; 13% were half-day, four-day week; and 4% of the programs had mixed schedules. The average class size was 13 children.

Children

In order to participate in the study, children in the participating classrooms had to meet three eligibility criteria: (1) the child was at least four years old by August 1 of the preschool year, (2) the child was intended to enter kindergarten in the following year, and (3) the caregivers provided informed consent. To identify eligible participants, backpack mails were sent home with all children in classrooms where teacher consent had been obtained. Caregivers were given three weeks and three opportunities to review study information and make informed decisions about whether to provide consent. Of all consents collected, children who met the forementioned eligibility criteria (1) and (2) were enrolled.

Prior to the start of the school year, participating classrooms were divided into two groups: business as usual (BAU) and kindergarten transition practices intervention (KTP), using a random assignment approach. Within the KTP classrooms, participants were further randomly assigned to receive only the classroom-based intervention activities or the supplemental home-visiting program (KTP +), resulting in a total of 119 children across 26 classrooms participating in home-visiting program in the first cohort of the study. After accounting for attrition that occurred prior to the onset of intervention, the analytical sample for the current study involving home-visiting participating patterns includes 113 families from 26 classrooms.

Table 1 provides an overview of the analytical sample: 53% were boys, 14% were receiving special-education services based on receipt of an Individualized Education Plan (IEP), and 3% resided in non-English-speaking households. On average, children were 55 months old (with a range of 48 to 66 months), and there was a diverse racial and ethnic composition (41% non-Hispanic Black, 34% non-Hispanic White, 10% Hispanic, and 15% other or multiracial). The sample is characterized as very low SES, with a median annual household income ranging from $20,000 to $30,000. Specifically, 45% of caregivers reported an annual household income of less than $20,000, 31% reported an income of $20,000 to $40,000, and only 24% had annual household incomes exceeding $40,000. Moreover, slightly more than one-half (55%) of the children’s mothers had a high school diploma or equivalent as their highest educational credential, while 18% had an associate degree, and 14% had a bachelor’s degree or above. The remaining 13% had the highest education credential of less than high school. On average, households had two to three children (M = 2.65, range = 1 ~ 8) and one to two adults (M = 1.67, range = 0 ~ 4), although 53% of the children lived in households with a single caregiver.

Table 1 Descriptive statistics of the analytical sample: families receiving home visiting program
Full size table

General Procedures

In this study, the focus was on classrooms in which the KTP + intervention was implemented, which involved parents participating in a home-visiting program to augment kindergarten-transition practices implemented in the classroom. Specifically, in classrooms assigned to the KTP condition, teachers implemented a range of practices to support children’s transition into kindergarten over the academic year of preschool. These practices were aligned to those described in the manual Successful Kindergarten Transition (Pianta & Kraft-Sayre, 2003). Classroom-based intervention activities were facilitated by five transition coordinators, each of whom was assigned to approximately five classrooms. The classroom activities included, for instance, periodic mailings to parents regarding the importance of preparing for the kindergarten transition, small-group lessons for children about typical kindergarten classroom rules and routines, and visits with kindergarten students and teachers.

Children who were randomized to the KTP + condition (from among classrooms assigned to KTP) received up to nine intended home visits over approximately 15 months, beginning at the start of preschool and extending through middle of kindergarten year. The timeline and detailed home-visiting activities are presented in Table 2. The family’s transition coordinator scheduled visits via the parents’ preferred method of contact, and each visit lasted around 45 min. Caregivers were sent reminders prior to each visit in an effort to maintain participation.

Table 2 Home visiting schedule and focus for the pre-k and kindergarten years
Full size table

The first two home visits focused on transition coordinators establishing rapport with caregivers and completing a needs assessment to develop a plan to support each family. Each remaining home visit was designed to address a specific goal as outlined in the Successful Kindergarten Transition manual, such as building home-school connections, and involved a home-learning activity with books and resources to promote children’s kindergarten readiness. Transition coordinators reestablished rapport at the start of each home visit, asking general questions to check in with how everything is going. Transition coordinators importantly noted two or three points of conversation to follow up on each visit, offering consistent support to the family. For example, transition coordinators would offer to facilitate communication with the child’s teacher, discuss strategies to better engage the child in at-home learning activities, and provide information about a potentially helpful topic or resource. Transition coordinators also used visits to discuss school-engagement opportunities and to encourage parent-school and parent-teacher connections. While families were encouraged to use the learning materials provided and attend in-person school events, transition coordinators were empathetic to any difficulties caregivers reported. Overall, transition coordinators were trained to adopt evidence-based practices which recognized families’ unique strengths and actively responded to their complex needs (Azzi-Lessing, 2011).

Measures

Measures of relevance to the present study included home-visiting measures, family and child characteristics, and children’s academic skills.

Home-Visiting Measures

The number of completed home visits by each family was the primary outcome variable in this study. We recorded the number of times the assigned transition coordinator successfully visited each family. Although there were nine planned home visits (see Table 2), transition coordinators sometimes had to cover multiple topics during one visit due to a missed prior visit. In such cases, the visit was counted only once. Cancelled or incomplete visits (e.g., due to caregivers being absent) were not counted. In addition, we recorded the time at which each family dropped out of the home-visiting program, after which the family stopped participating and did not complete any subsequent home visits. Based on this information, families were categorized as “retained” (maintained through the last home visit), “early dropout” (dropped out by the end of preschool year or before the fifth home visit), and “late dropout” (dropped out in the summer after preschool or later).

Child and Family Characteristics

Child and family characteristics were captured using caregiver surveys during the fall. Caregivers responded to questions about their child, themselves and their families, reporting information such as race/ethnicity, home language, mother highest level of education, annual household income, family size, among others. The list of child and family characteristics examined in this study is shown in Table 1.

Child Academic Skills

Children’s academic skills were assessed in fall and spring of the preschool year using three subtests of the Woodcock Johnson III NU Tests of Achievement (Woodcock et al., 2007): Letter Word Identification (WJ-LWID, early literacy), Applied Problems (WJ-AP, early numeracy), and Picture Vocabulary (WJ-PV, vocabulary). Children were individually administered these subtests by trained research staff in private settings in their school over the test window. For each subtest, starting items varied according to students’ age and the assessments followed basal and ceiling rules per the Examiner’s Manual (Mather & Woodcock, 2001). Responses were scored 1 if correct and 0 if incorrect, and then summed to create raw scores for each subtest.

For the current study, we converted raw scores to standard scores, which were age-adjusted and represented the relative position of a student among his or her peers in a nationally representative sample (national average = 100, SD = 15). In Table 1, we present the means and standard deviations of the WJ standard scores in fall of preschool to characterize the sample of children. Reliability for the current sample as measured by Cronbach’s alpha was 0.90 for reading, 0.85 for math, and 0.66 ~ 0.68 for vocabulary. The reliability coefficients for vocabulary is lower than that in the representative sample (0.81), potentially due to the high level of heterogeneity in early language development patterns within our low-income sample.

Analytical Approach

Descriptive Statistics, Correlational Analyses, and Between-Group Comparisons

A primary aim of this study was to investigate the retention of families in a home-visiting program aimed to enhance kindergarten readiness and to explore potential correlates of retention. The program encompassed nine visits over the span of approximately 15 months. We first conducted a descriptive analysis of the number of home visits completed and the time of drop-out for our sample of 113 families. Second, we calculated Pearson correlation coefficients to explore potential associations between the number of home visits and child or family characteristics. We further compared child and family characteristics across the three categories of families based on their drop-out status (i.e., retained, early drop-out, late drop-out) and tested for potential group differences using analysis of variance (ANOVA) for continuous variables and Chi-square tests for categorical variables.

Survival Analyses

In this study, we also applied survival analysis to investigate families’ retention status and duration of stay in the home-visiting program, as well as potential predictors of these outcomes. Survival analysis, initially developed for analyzing cancer data and death rates, is a statistical technique that can also be applied to studying other types of events. In this study, we use it to understand families' retention in a home-visiting program and their likelihood of drop** out. The event or “hazard” refers to families’ drop-out status, and “survival” refers to families’ retention during the duration of the home-visiting program.

We used two survival analysis models: life-table method and Cox proportional hazards model (Allison, 2010). Life-table method provides descriptive information of retention for each time-point, such as the number of families enrolled in the program before each visit (“Effective sample size”), number of families drop** out at each visit (“Number of dropouts”), the likelihood of drop** out (“Hazard function”) and the likelihood of retention (“Survival function”). To visualize the trends, we used Kaplan–Meier plots, which display how retention rates changed over time. Cox proportional hazards model, on the other hand, assumes that the risk of drop** out for any two individuals or subgroups is proportional over time. In other words, while the hazard of dropout may be different depending on individual characteristics, the ratio of the hazards remains constant over time. Cox proportional hazards model allows us to examine the relationship between the log likelihood of drop** out and each child- and family-level covariate.

Missing Data

In this study, the key outcome variables including the number of home visits, time of drop-out, and completion status were fully observed. There was a range of missing data for child and family characteristics, ranging from 0 to 33%, and 5% for child academic measures. For the survival analyses, listwise deletion was used resulting in 0% of missing data for the hazard model and 10% for the Cox proportional model.

Results

Home Visiting Participation for Families During the Preschool-to-Kindergarten Transition

Figure 1 presents the number of home visits completed by the 113 families who participated in the KTP + intervention. On average, families completed four out of the nine planned home visits with large dispersion in the number of visits completed (SD = 3.28, range = 0 ~ 9). These data reveal that 13% of families did not complete any home visits, whereas 38% completed fewer than five visits, and only 17% completed all nine home visits as originally planned.

Fig. 1
figure 1

Total Number of Home Visits Completed (upper) and Percentage of Families That Completed Each Home Visit (lower). HV 1: Nov-Dec 2018; HV 2: Jan-Feb 2019; HV 3: Feb-May 2019; HV 4: May-Jun 2019; HV 5: Jun-Jul 2019; HV 6: Jul-Aug 2019; HV 7: Sept-Oct 2019; HV 8: Oct-Nov 2019; HV 9: Dec 2019-Jan 2020

Full size image

In terms of the timing of dropout, 61% of the families dropped out before the final home visit, indicating a high rate of attrition. As shown in Fig. 2, approximately one-third of all participating families dropped out during the preschool year, before the summer transition that corresponded to home visit 5. Another 13% of the families were lost during summer (visits 5 and 6), while 18% dropped out during the kindergarten year.

Fig. 2
figure 2

Distribution of the Time at Which Families Dropped Out of the Home Visit Program. HV 1: Nov-Dec 2018; HV 2: Jan-Feb 2019; HV 3: Feb-May 2019; HV 4: May-Jun 2019; HV 5: Jun-Jul 2019; HV 6: Jul-Aug 2019; HV 7: Sept-Oct 2019; HV 8: Oct-Nov 2019; HV 9: Dec 2019-Jan 2020

Full size image

Additionally, we examined the percentage of families who completed each of the nine home visits within its respective time window (Fig. 1, lower pane). The results indicate that the completion rate was approximately 70% for the first two visits, with a gradual decline to 42% during the summer between preschool and kindergarten, which corresponded to home visit 5. Roughly half of the sample completed the first home visit of the kindergarten year, which was home visit 6, while only 39% of caregivers maintained their participation through to the ninth and final visit.

Correlates of the Number of Completed Home Visits

We investigated the correlation between the number of home visits completed and child-family characteristics. Our analysis found that the number of home visits completed was not associated with child gender, ethnicity (Hispanic vs. non-Hispanic), home language (English-speaking household vs. other), child IEP status, or family structure variables (such as the number of adults or children in the household, or whether the family had a single caregiver). We also found no correlation between the number of visits completed and children’s numeracy or vocabulary skills. However, we found that families with older children (r = -0.23) or whose children were Black (r = -0.30) completed fewer home visits, whereas families with college-educated mothers (r = 0.23) and those in higher income quartiles (r = 0.21) completed a higher number of visits. The child’s baseline literacy skills exhibited a positive correlation with both the number of home visits completed and the length of program participation (r = 0.19).

Comparing Retained Families, Early Drop-Out and Late Drop-Out

Previously noted, we categorized the participating families into three groups based on their retention in the home-visiting program: retained (N = 44), early drop-out (i.e., families lost during the preschool year, N = 39), and late drop-out (i.e., families lost during or after the summer before kindergarten, N = 30). Table 3 provides a comparison of child and family characteristics across these three groups, using descriptive statistics and statistical tests. Note that we did not adjust for multiple comparisons, so the test results should be interpreted with caution and viewed as exploratory. Our analysis indicated that families with Black children were more likely to drop out of the home-visiting program compared to families of other racial backgrounds (p = 0.026). Specifically, 77% of Black families dropped out before the end of the program, compared to 52% of non-Black families. Additionally, families who dropped out tended to be in lower income quartiles than those who retained (p = 0.014). Furthermore, families with children who scored higher in baseline literacy tended to drop out later or to be retained in the program, whereas families with children who scored lower in baseline literacy tended to drop out early (p = 0.039).

Table 3 Comparisons between families with different levels of participation
Full size table

Survival Analyses of Retention in Home Visiting: Patterns and Predictors

To investigate the retention rates over time in the home-visiting program, we employed survival analysis. The nonparametric life-table survival estimates are summarized in Table 4, and the corresponding survival curve is presented in Fig. 3. Our results indicated that families who dropped out were most likely to do so at the beginning of the program, with the highest chance of dropouts (13%) occurring at time-point 0, before the first home visit. The probability of family retention at the beginning of the program was 87%. At the end of the program, corresponding to home visit 9, the probability of family retention decreased to 39%.

Table 4 The life-table survival estimates of family dropout
Full size table
Fig. 3
figure 3

The Kaplan–Meier Plot of Family Dropout. The grey shade represents the 95% confidence interval

Full size image

To examine the predictors of dropout (i.e., “hazard” in the survival analyses) in the home-visiting program, we conducted a Cox proportional hazard model that included selected child and family characteristics identified as significant correlates of program participation as reported in the previous sections. We tested these variables simultaneously while controlling for the effects of transition coordinators. As shown in Table 5, child age had a significant positive relationship with the probability of drop-out (B = 0.10, Exp(B) = 1.11, p = 0.009), indicating that with each additional month of age, the hazard of drop** out increased by 11%. Furthermore, we found that an increase of one point on the standardized academic skills scale at baseline was related to 2% lower hazard of drop** out (B = -0.02, Exp(B) = 0.98, p = 0.087).

Table 5 The cox proportional hazard model predicting the probability of family dropout (N = 102)
Full size table

Discussion

Home-visiting programs represent an important investment in young children and families in terms of targeting a variety of early child- and family-level outcomes. Although some evidence supports the effectiveness of these programs, studies generally show small effect sizes (approximately 0.10, on average) (Supplee and Dugan, 2019) and the evidence of program effectiveness is mixed, with some studies reporting positive outcomes (e.g., Bugental et al., 2002; Lee et al., 2009) whereas others report non-significance (e.g., Barth, 1991; Duggan et al., 2004; Henwood et al., 2020; Kartin et al., 2002; Michalopoulos et al., 2019). Family attrition and lack of full participation may be one factor contributing to these small effect sizes and mixed findings (Janczewski et al., 2019); for instance, a subset of families may not be exposed to the full and intended strength of a program’s activities. In part, this may occur because home-visiting programs last a long period of time to provide sustained support to families (Nievar et al., 2010), and thus require a long and sustained commitment. In previous work, researchers associated a variety of explanations for family attrition from home visiting programs (Fifolt et al., 2017). These include: (a) lack of perceived benefits (Holland et al., 2014a, 2011), especially when partnering with low-income families (Roggman et al., 2008), suggests the need to rigorously identify ways to promote participation and long-term maintenance of families in home-visiting programs targeting school readiness. Given the nature of the KTP sample, which largely involved very low-income families, it is important to acknowledge the systematic challenges that these families face as they navigate competing obligations in order to ensure access to necessary, and potentially scarce, resources with the little time they have (Beasley et al., 2018; Roy et al., 2004). In alignment with the ecological perspective of KTP, the overarching goal of the home-visits examined in the present study was to build supportive connections among families, teachers, schools, peers, the community, and of course, the transitioning child. However, families may feel the need to prioritize other, more basic goals like the child’s nutrition or behavior problems, or have other concerns about their child’s development and school readiness such as their early literacy or math skills. If the specific needs, values, and goals of families are not addressed through flexible implementation of home visiting programs, retention is likely to be problematic (Azzi-Lessing, 2013). In the present study, although family needs were assessed during the first home visit to inform the provision of future visits, attrition patterns suggest that it is critical to additionally attend to programmatic factors such as design and content.

Implementation-science frameworks, including the Theoretical Domains Framework (TDF; Michie et al., 2008), offer an important advancement for empirically identifying barriers that prevent individuals from accessing empirically based programs and practices. The TDF provides a systematic approach to identifying barriers individuals face, and map** evidence-based behavior-change strategies to these barriers as a way for promoting access to and participation in programs and practices. For instance, in an application of the TDF to home-based literacy-promotion strategies for parents with young children, Justice et al. (2018) tested four behavior-change techniques for their impacts on parents’ program implementation: rewards, encouragement, feedback, and modeling. This study found that rewards and encouragement significantly influenced parents’ program implementation, which in turn improved child outcomes. Applications of this approach to identifying strategies to promoted caregiver maintenance in home-visiting programs may identify approaches that not only maintain caregiver and child participation in school-readiness interventions over time, but also improve targeted child outcomes, such as mitigating kindergarten-transition difficulties.

The second major finding of this study was that loss of families occurred across the duration home-visiting period. Interestingly,13% of families attrited prior to program commencement and did not receive any home-visiting activities. Other studies involving parents of preschool-aged children in longitudinal early-learning interventions have also found relatively high rates of attrition at program commencement. For instance, in implementing a 15-week home shared reading program for children with disabilities, 20% of caregivers left before starting the program (Justice et al., 2020). In another study involving a year-long home shared reading program, also for children with disabilities, 38% of caregivers were classified as ‘early drop-outs;’ these caregivers dropped out by the 5th (out of 30) week of the program, with the majority attriting prior to program start (Logan et al., 2019). Of relevance to the present findings, caregivers who were classified as early dropouts tended to be significantly less educated than those caregivers who completed the 30-week program, with a medium-to-large effect size (d = 0.5; calculations our own).

It is not entirely clear why caregivers may enroll in an early-learning intervention for their children and subsequently withdraw before starting program activities, but the pattern of early loss appears relatively consistently across cited works. This pattern may reflect the social desirability of program enrollment, in that participants enroll in programs that are viewed as desirable for them and their children. Or, this pattern may reflect a gap between one’s intentions and future behaviors, an oft-cited phenomenon within the health-sciences literature (Conner et al., 2016), or the effect of unanticipated future constraints in conducting program activities. Again, an implementation-science lens, such as the TDF, is likely to be informative to determining the basis for early drop-out in home-visiting interventions and strategies to address this phenomenon.

Not all home-visiting participants attrited at study start, with about one-third (35%) of participating families lost during the preschool year and another surge of dropout (13%) occurred during the fifth and sixth home visits during the summer break following the preschool academic year. This pattern of dispersion in participant loss is similar to that reported by Logan et al (2019) when implementing a home book-reading program with parents of preschool children over an academic year. As past work suggests (Beasley et al., 2018), fostering both initial and sustained engagement is a crucial, albeit challenging, aspect of longer-term interventions. Though our home-visiting program intended to support children and families in their transition into kindergarten, most attrition occurred prior to this transition, thus potentially lessening potential program effectiveness. Building upon archival research focused on attrition within other prevalent home-visiting programs will be instrumental in develo**, testing, and scaling strategies to maintain caregivers in long-term home-visiting programs targeting successful kindergarten transitions.

The third major finding was that retention in the home-visiting program was not equitable across families based on child race, SES-related variables of maternal education and household income, and children’s baseline literacy skills. First, in consideration of racial inequities in family participation in the home-visiting program, study findings showed that 77% of self-identifying Black families dropped out by the final home visit compared to 52% of non-Black families. This significant difference in family maintenance as a function of race potentially implicates the presence of structural racism and the consequential discriminatory practices, which often culminate in a mistrust of powerful institutions among minoritized families (Bazargan et al., 2021; Braveman et al., 2022). Of note is that the home-visiting program was branded as a federally funded study led by a large state institution of higher education, and KTP coordinators were predominantly white educated women. Community-based programs delivered by community residents and featuring tailored and culturally relevant programming may serve as a mechanism to address racial disparities in family maintenance in long-term kindergarten-readiness initiatives (Two Feathers et al., 2005). For example, the books and activities provided in the KTP program could have been more intentionally selected after district recruitment to better align with families’ ethnicities, experiences, and cultural practices (Rodríguez, 2014). Likewise, staffing patterns in such programs may also provide a key mechanism in maintaining participation, as shown in a study by Daro et al. (2003) that used a method of intentionally matching participants with home visitors in terms of ethnic identity and other characteristics. This resulted in better retention rates for African American families compared to Hispanic and white families. The present study did not match home visitors and family characteristics and was limited to assigning home visitors to families based upon geographic proximity.

A paradoxical finding of the present study also concerned the positive associations between family SES (maternal education and household income) and child initial skill levels in terms of program maintenance. That is, maintenance in study activities occurred among those caregivers and children who would benefit the least, consistent with a rich-get-richer Matthew Effect (Coyne et al., 2019). This finding suggests that parents who are more highly educated, who have more resources, and who have more highly skilled children were more likely to continue in the home-visiting program and, presumably, benefit from this offering. Other studies of home-visiting programs have also reported a rich-get-richer phenomenon. For instance, Haroz et al. (2019) examined the effects of the Family Spirit home-visiting program designed for Native American mothers with infants, finding that benefits were most significant for (relatively) more educated mothers and families with more stable housing. This is also consistent with findings in literatures studying treatment completion in pediatric programs, where children who experienced more symptoms and caregivers who reported higher levels of stress were more likely to prematurely drop out (e.g., Pellerin et al., 2010). The paradoxical consequence of this Matthew Effect phenomenon observed in home-visiting programs, such as KTP + , is that these can further enlarge the gap among children in their development, including school readiness.

Limitations and Future Directions

Several limitations warrant discussion and can serve to identify directions for future research. First, the present study consisted of a relatively small sample size (N = 113) of low-income families in an urban region in the Midwest. These observed retention and attrition patterns reported here may not generalize to other populations and contexts, such as families living in rural areas or those considered middle- or upper-class. As there are relatively few studies focused on home-visiting programs targeting school readiness, studies involving diverse populations are needed. Second, although this study identifies patterns of attrition, there was no mechanism for knowing exactly why parents withdrew from home visit program involvement. That is, we did not systematically identify the barriers that impeded ongoing participation in the home-visiting program. Future work (e.g., Fifolt et al., 2017; Wagner et al., 2003) would benefit from collecting information on families’ rationale for attrition and barriers to engagement, which would be informative to applications of implementation-science paradigms to reduce barriers. Finally, the present study was conducted prior to the global pandemic, the conditions of which have substantially affected families and children in myriad ways. For instance, family routines have been modified (Singletary et al., 2022) and children’s school readiness has been negatively affected (González et al., 2022). As with all forthcoming research on educational interventions, it will be necessary to determine whether study findings generalize to the present era.

Conclusion

The transition from preschool to kindergarten is an important and challenging milestone for many children and their families. How children navigate this transition has implications for their later school success. Home-visiting programs can support the development of children’s cognitive and socio-emotional skills, easing this transition to kindergarten and supporting future learning, especially for children experiencing risk factors that compromise school readiness. The patterns of caregiver attrition identified in this home-visiting program designed to promote kindergarten readiness in a sample of low-income families indicated problems with retention from the outset, which later varied by child age, race, and skills. While more work is needed, the present study highlights initial implementation challenges and the sustainability of a home-visiting program. The findings additionally emphasize the need to examine how the provision of differential supports in home-visiting programs may achieve long-term retention of those vulnerable to attrition.