1 BACKGROUND

Evidence from previous research in early human development demonstrates the importance of the first 1000 days from conception to positive and life-long child outcomes (Britto et al., 2017; Cusick & Georgieff, 2016; Kraemer et al., 2018). This body of research calls for evidence-based early intervention services within this sensitive window of time, targeting parents, caregivers, and children who have been identified as at risk of poor outcomes (Daelmans et al., 2017). Home visiting has been considered a promising strategy for addressing multiple needs of families, especially those who are experiencing adversity (Avellar & Supplee, 2013; Filene et al., 2013; D. McNaughton, 1994; Molloy et al., 2021; Peacock et al., 2013).

The Maternal Early Childhood Sustained Home-visiting (MECSH) program is a structured nurse facilitated program designed to address health inequities by catering for families experiencing adversity (Kemp et al., 2011, 2017). The MECSH program is delivered to families with significant risk factors by university-trained registered nurses (baccalaureate) with postgraduate training in child and family health nursing. The nurses were also provided additional online and face-to-face training in the MECSH program as well as reflective practice supervision. The program commences in pregnancy and continues until the child is 2 years old. The program's goals are to: improve the transition to parenting by supporting mothers through pregnancy; improve maternal health and well-being by helping mothers to care for themselves; improve child health and development by helping parents to interact with their children in developmentally supportive ways; develop and promote parents’ aspirations for themselves and their children; and improve family and social relationships and networks by helping parents to foster supportive relationships within the family and with other families and services (Kemp et al., 2011).

Previous research evidence has shown MECSH-based programs to be effective, reporting significant improvements in maternal confidence in care, knowledge and experience, positive child health and development outcomes, and creating positive home environments to support healthy child development (Goldfeld et al., 2018, 2019; Kemp et al., 2011, 2018). However, there is limited knowledge in relation to the specific mechanisms of effective practice that achieve high quality of care in home visiting interventions. It has been previously noted by Kemp (2016, p. 429) that “quality is achieved by identifying and measuring the core ingredients and variations.” However, adaptations and variations in the practices of home visiting programs have not been investigated (Roggman et al., 2001, 2016). Variations in the care provided by home visiting nurses to families, how interventions are structured, and the quality of interactions during the visits may determine how effective a home visiting program is for a particular family (Filene et al., 2013; Nygren et al., 2018; Roggman et al., 2001).

Little research has considered which practices and processes of home visiting contribute to positive outcomes and for whom. The quality of implementation is often poorly reported and, when reported, describes what happens on home visits in very general ways (Kemp et al., 2019; Roggman et al., 2016). Researchers have concluded that what specifically occurs during a home visit is largely unknown and remains a “black box” (Goldfeld et al., 2018; D. B. McNaughton, 2004).

Variations in care occur for a range of different reasons (Australian Commission on Safety & Quality in Health Care, 2015). When a variation is desirable and warranted to customize or tailor programs to meet the client's unique needs and preferences (Kreuter & Skinner, 2000), the variation can be considered “purposeful” variability. Home visiting care services provide opportunities for nurses to observe the environment in which families live, which can help them identify a family's unique needs and provide a greater level of individual attention than usual facility-based care (Goldfeld et al., 2018). This variation is purposefully and intuitively created by nurses based on their knowledge, skills, observations, communications, and relationships with clients. This kind of practice variation is compatible with concepts such as individualized care (Suhonen et al., 2002, 2005, 2007, 2008, 2012; Wright & McCormack, 2001), tailored-care (Pasick, 1997), client-centered care (Brown et al., 2006), and people-centered care (Lydahl, 2021; World Health Organization, 2007). Bespoke care models are largely preferred by clients and thus result in better outcomes and higher rates of client satisfaction (Bertakis & Azari, 2011; Ekman et al., 2012; Jo Delaney, 2018).

On the other hand, if a variation in the care delivery model is unwarranted, it may signal that clients are not receiving appropriate care (Australian Commission on Safety & Quality in Health Care, 2015). Lack of compliance with evidence-based program elements can result in serious consequences, including drift and dilution (Kalisch et al., 2011; Kemp, 2020). Drift is defined as “a misapplication or mistaken application of the model, often involving either technical error, abandonment of core and requisite components, or introduction of counterproductive elements” (Aarons et al., 2012). Dilution is the failure to deliver the intensity or duration of the program as intended (Goldfeld et al., 2018). Non-purposeful variability can threaten the fidelity of the program and the quality of the care provided, while purposeful variability may enhance it.

Achieving the appropriate balance between program compliance and purposeful variation to meet individual needs underpins the theoretical concept of precision home visiting. Precision home visiting is home visiting that differentiates what works, for whom, and in what contexts to achieve specific outcomes (Home Visiting Applied Research, Collaborative; Supplee & Duggan, 2019). It focuses on the components of home visiting services that are most likely to be effective in light of mothers’ and families’ characteristics and social and cultural context (Haroz et al., 2019). Mothers participating in home visiting programs are, in general, provided with the same program content, dose and duration regardless of their needs or circumstances. However, mothers and children may not need the same program content or dose. Furthermore, without clear guidance on how to customize service delivery, home visiting nurses’ judgement may compromise program fidelity. Thus, there is an emerging demand for new home-visiting strategies to address the diverse and critical needs of mothers and families while maintaining fidelity to the core ingredients of evidence-based home visiting programs (Haroz et al., 2019).

This study builds upon and extends the existing home visiting research literature by exploring variations in care. It aimed to examine the variations in care for mothers and children in the delivery of the right@home MECSH-based program by answering four research questions: What are the variations in care as per the program schedule and requirements in the delivery of the program? (Compliance); How did care vary over the duration of the program? Which variations in care in the delivery of the program are made in response to the families’ individual risks? (Customization according to families’ risk factors: purposeful variability); and How precisely did the program content vary in response to families’ risk?

2 METHODS 2.1 Study design

This study draws on data collected in a randomized controlled trial (RCT) of right@home, a sustained nurse home visiting program from pregnancy to child age 2 years (Goldfeld et al., 2017). It was a MECSH-based program which was trialed in seven localities in the Australian states of Victoria and Tasmania (Goldfeld et al., 2017).

2.2 Participants

Eligibility criteria for the right@home trial included pregnant women attending the antenatal clinics in Victoria and Tasmania from May 2013 to August 2014, who were less than 37 weeks gestation, had sufficient English proficiency to verbally answer interview questions, resided within the study travel boundaries, and reported two or more of ten sociodemographic risk factors for adverse parent and/or child outcomes in risk factor screening conducted verbally by trained research assistants working in antenatal clinics: young pregnancy (age <23 years); not living with another adult; no support in pregnancy; smoking; poor/fair/good health; long-term illness; anxious mood; not completed Year 12 secondary level education; no income; and never worked (Goldfeld et al., 2017; A. Price et al., 2019; A. M. Price et al., 2017). The right@home trial recruited 722 pregnant women, including 363 women in the intervention group and 359 women in the control group. The control group received usual care (Goldfeld et al., 2017).

2.3 Measures 2.3.1 Participants sociodemographic and risk factors

Data collected at women's commencement in the program, including sociodemographic and risk factor screening data, and children's dates of birth were extracted from the trial enrolment data.

2.3.2 Visit content provided for mothers and families

As part of the study, visit content was recorded by the nurses at the completion of each visit to the woman and her family (Goldfeld et al., 2018; Kemp et al., 2019). The nurse completed an online checklist designed explicitly for use in the program quality monitoring. The checklist included the unique client identifier, date of the visit and the activities undertaken. The electronic checklist was located on the nurse's mobile device (tablet) with a simple touch entry. It was used to record activities and content provided in the visit. The checklist identified activities or topics discussed with the family across nine headings: infant well-being; maternal well-being; maternal mental health; family well-being; preventive health care; environment/resources; planning and goal setting; referrals; and tools and focus modules. There were 48 items in the antenatal checklist and 56 items in the postnatal checklist (Goldfeld et al., 2018; Kemp et al., 2019).

2.4 Ethics approval

The right@home trial was approved by the Human Research Ethics Committees in Australia of the Royal Children's Hospital, Victoria (HREC 32296), Peninsula Health, Victoria (HREC/13/PH/14), Ballarat Health Services, Victoria (HREC/13/BHSSJOG/9), and The University of Tasmania (HREC H0013113). All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethics standards. Written consent was obtained from all participants.

2.5 Analytic strategy

Data analyses were performed using R (4.0.5) and RStudio (Version 1.4.1106). Descriptive statistics, including numbers and percentages, were computed for the variables involved in this study. Odds ratios (OR), accuracy, precision, sensitivity, F1-score, and Matthews Correlation Coefficient (MCC) were calculated to analyze the risk-related variability in care. The F1-score is the harmonic mean of the precision and sensitivity, where the score reaches its best value at 1 (perfect precision and sensitivity) and its worst value at 0 (Choi et al., 2020). Several studies show that MCC produces a more informative and truthful score than accuracy and F1-score in evaluating binary classifications (Boughorbel et al., 2017; Chicco & Jurman, 2020; Chicco et al., 2021). MCC ranges from −1 and 1, with −1 indicating perfect negative correlation, 0 random distribution, and 1 perfect correlation (Boughorbel et al., 2017; Vihinen, 2012).

There were six focus modules (required curriculum) for the right@home program, scheduled to be delivered at specific times related to the developmental needs of the child (Goldfeld et al., 2018, 2019): Get Up & Grow; Sleep program; Safety audit; Promoting First Relationship; Video feedback; and Learning to Communicate. For the analyses on the variations in care content provision, the required visit contents for each of five focus modules were identified according to the right@home intervention manual (Table 1). The module of Video feedback was not included in the analysis as it was part of the Promoting First Relationship and Learning to Communicate modules.

TABLE 1. Visit content applicable to the five focus modules Focus module Visit content Get up & grow Infant health and growth, infant development, infant feeding, personal health record complete, get up and grow healthy eating guide Sleep program Infant crying, infant sleeping and settling, SIDS prevention Safety audit Infant bathing, SIDS, household safety, car safety Promoting first relationships Infant development, interaction between parent and infant, Promoting First Relationship tools, video feedback Learning to communicate Infant health and growth, infant development, infant crying, infant feeding, infant bathing, video feedback, Learning to Communicate and small talk tools

The visit content that could be used to identify the delivery of each focus module was mapped out according to the right@home implementation manual. Using this content mapping, the complete set of home visit activity records were scanned to identify if the visit content was delivered either at the scheduled time or within an acceptable time frame as per the home visit schedule. The time period between each visit of interest and the scheduled age of delivery was calculated, and the visit closest to the target child age was determined and marked. Content provision was assessed for each of these marked home visits. The service was considered to have been provided if any of the selected content was completed during the home visit, and the home visit occurred within the “on time” frame. We considered that the content was provided “within an acceptable time frame,” if the service was provided at least once between one scheduled visit before and two scheduled visits after the target visit. The service was considered to have been provided at a time appropriate to meet the child's developmental needs with the recognition that variation in child development may have required the nurse to deliver content somewhat earlier or later than scheduled. If these conditions were not met, the home visit was considered as “content not provided.”

3 RESULTS

Of 363 women randomized to the intervention group, 352 women (97.0%) commenced the intervention, and 304 of these women (86.4%) completed the right@home program when the child reached 2 years of age.

3.1 Variations against the planned/scheduled content

Variations in care were analyzed from the viewpoint of compliance with the scheduled content. For almost all five focus modules, the majority of the mothers were provided with scheduled content on time or within an acceptable time range, except for the sleep program module. Only 58.4% of women received at least one content of the sleep program module at the scheduled point of the antenatal period.

According to the program schedule, content for the Get Up and Grow module was required to be provided during the antenatal period, and child age 2, 4, 19, 52, and 104 weeks. More than 90% of women received scheduled content on time at almost all the required visits scheduled, except for the antenatal visit (60.1%). The sleep program module was provided at child age 26 weeks on time for 89.7% of participants, and within an acceptable time range for 98.1%. For the safety audit module, 79.2% of the participants received the content on time at the point of child age 3 weeks. However, the proportion increased to 97.3% when the provision of scheduled content within an acceptable time range was calculated. At least 98% of the participants received the scheduled content of the Promoting First Relationships and Learn to Communicate module within an acceptable time range, as presented in Table 2.

TABLE 2. Provision of scheduled visit content for five focus modules (N = 352) Provision of scheduled content “on time” Provision of scheduled content “within an acceptable time range” Required schedule for each module Total number of families Number of families provided content % Number of families provided content % “Get Up and Grow” Antenatal 296 178 60.1 246 83.1 2 weeks 339 316 93.2 338 99.7 4 weeks 339 314 93.2 339 100.0 19 weeks 328 297 90.5 326 99.4 52 weeks 305 290 95.1 305 100.0 104 weeks 275 266 96.7 275 100.0 Sleep program Antenatal 296 112 37.8 173 58.4 26 weeks 314 287 89.7 314 98.1 Safety audit Antenatal 296 222 75.0 275 92.9 3 weeks 336 266 79.2 327 97.3 38 weeks 312 267 85.6 305 97.4 Promoting first relationships 2 weeks 339 310 91.4 335 98.1 4 weeks 339 306 90.3 336 99.1 5 weeks 338 307 90.8 337 99.7 10 weeks 331 307 92.7 330 99.7 15 weeks 330 307 93.0 329 99.7 45 weeks 310 290 93.5 310 99.7 61 weeks 301 278 92.4 299 99.3 70 weeks 294 274 93.2 293 99.7 87 weeks 282 276 97.9 282 100.0 96 weeks 277 252 91.0 278 100.0 104 weeks 275 265 96.4 275 100.0 Learning to communicate 3 weeks 336 317 94.3 336 100.0 6 weeks 337 304 90.2 337 100.0 10 weeks 331 318 96.1 330 99.7 15 weeks 330 314 95.2 328 99.4 19 weeks 328 298 90.9 327 99.7 22 weeks 322 297 92.2 322 100.0 26 weeks 320 306 95.6 319 99.7 32 weeks 316 303 95.9 314 99.4 38 weeks 312 297 95.2 313 100.0 45 weeks 310 297 95.8 311 100.0 52 weeks 305 291 95.4 305 100.0 3.2 Variability in content delivery by different time frames

The proportion of the participants who received each visit content once or more was calculated by the different time frames: antenatal to child age 6 months, child age 7–12 months, and child age 13 months to 2 years. The most frequently provided visit content for three different time frames were “mental health” (99.7%) for antenatal to child age 6 months, “infant health and growth” and “infant development” (100% and 99.7%, respectively) for the period of child age 7–12 months as well as the period of child age 13 months to 2 years. The least provided visit contents were “family law” (24.7%) for antenatal to child age 6 months, “Edinburgh Depression Scale” (12.4%) for child age 7–12 months, and “sterilisation for feeding” (6.7%) for child age 13 months to 2 years.

The content for which provision varied by different timeframe (more than 40% difference) were: “sterilization”; “infant bathing”; “parent craft”; “pregnancy and childbirth”; “maternal smoking”; “expectations and reality of having a baby”; “drugs and alcohol”; “car safety”; “Edinburgh depression scale tool”; and “Learning to Communicate tool” (See Additional table).

Additional table. Number and percentage of mothers who received the visit content at least once during the indicated timeframe Content Antenatal to child 6 months (n = 352) Child 7–12 months (n = 314) 1–2 years (n = 300) n % n % n % Parent craft/Infant well-being Infant health and growth 343 97.4 314 100.0 299 99.7 Infant development 343 97.4 314 100.0 299 99.7 Interaction between parent and infant 343 97.4 311 99.0 298 99.3 Infant crying 340 96.6 224 71.3 193 64.3 Infant sleeping/settling 347 98.6 312 99.4 295 98.3 Infant feeding 349 99.1 313 99.7 297 99.0 Sterilization 228 64.8 61 19.4 20 6.7 Infant bathing 252 71.6 87 27.7 71 23.7 Parent craft (clothing/nappies, etc) 297 84.4 116 36.9 114 38.0 Maternal well-being Maternal Health 351 99.7 313 99.7 298 99.3 Physical activity 339 96.3 256 81.5 277 92.3 Maternal nutrition 342 97.2 239 76.1 256 85.3 Dental care 152 43.2 64 20.4 104 34.7 Sexual activity 180 51.1 83 26.4 85 28.3 Pregnancy Childbirth 221 62.8 0 0.0 0 0.0 Contraception/conception 292 83.0 144 45.9 146 48.7 Maternal smoking 281 79.8 96 30.6 115