ORIGINS comprises non-active participants (n=4,457 mothers, 5,227 children and 1,117 non-birthing partners) and active participants (3,448 mothers, 3,806 children and 1,403 non-birthing partners), who are distinguished by the data provided by each. Non-active participants are those who were willing to be involved in the study, but did not want to commit to the extra time and effort required to be an active participant. Instead, they consented to ORIGINS linking to data routinely collected by hospital and midwives, antenatally and during birth. For active participants, as well as linking to this routine data, additional data are collected by ORIGINS over several timepoints, antenatally to five years of age, via online surveys, face-to-face clinical appointments and the collection of biosamples. Both active and non-active participants could then consent to ORIGINS linking to additional State and Commonwealth Government data.

Hence, ORIGINS consent forms were broken into three parts:

Note that, unless otherwise specified, we hereafter use the term ‘participant’ to refer to all types of participants, both active and non-active, and mothers, non-birthing partners and children. While all active participants have now been recruited (active participants were recruited from 2017 to 2023), recruitment of non-active participants is ongoing, and should be complete by 2024, having started in 2017 as well. In addition, ORIGINS aims to enhance the databank through external linkage to available State and Commonwealth datasets (e.g. Medicare, the National Assessment Program – Literacy and Numeracy, the Pharmaceutical Benefits Scheme, and the Australian Immunisation Register).

Once consented into The ORIGINS Project, each individual is issued an ORIGINS Unique Identification Number (OUID), including the mother, non-birthing partner and children. If the mother or non-birthing partner have previously consented into the study with a different pregnancy, they will retain the previously issued OUID. This allows ORIGINS to identify siblings within the databank. An ORIGINS Pregnancy Number (OPN) is generated for each pregnancy, which is especially important for multiple births (i.e. twins and triplets). The JHC Medical Reference Number (MRN) is collected from Meditech (the database used by JHC) and entered into The ORIGINS Project database. Figure 1 provides an overview of participant recruitment and retention across the antenatal and birth timepoints. The top half of the figure shows the number of pregnancies consented to the study antenatally, whereas the bottom half of the figure shows the mothers and children who remain consented in the study after birth. Of the 17,403 women approached across one or more pregnancies antenatally or at birth, 8,043 (46.2%) declined, 3,448 (19.8%) remained consented as active participants at birth (with 3,806 children), and 4,457 (25.2%) consented to be non-active participants at birth (with 5,227 children).

Women giving birth at JHC tend to be primarily from Joondalup, north of Perth, the capital city of Western Australia. Table 1 shows the demographic characteristics of Joondalup city compared to Western Australia and Australia [12]; a more detailed comparison can be found at 2021 Joondalup, Census All persons QuickStats | Australian Bureau of Statistics (abs.gov.au). Note that, these data were not available for mothers in Joondalup specifically, and ORIGINS collects data about families, so the demographic characteristics of Joondalup generally have been provided. The table shows that, compared to Western Australia and Australia, fewer people in Joondalup identify as Aboriginal and/or Torres Strait Islanders, are more likely to be married, and tend to be slightly better educated, are more likely to be in “white-collar” professions, from North Western Europe and identify as Christian or with no religion.

Table 2 provides a comparison of key demographic variables among active versus non-active mothers. The table shows that, compared to non-active mothers, active mothers are, on average, slightly older (less than half a year older), are more likely to be having their first baby, have a higher mean socioeconomic status (0.2 deciles, or 2%, greater) as indicated by their higher average Index of Relative Socioeconomic Advantage and Disadvantage, and are less likely to identify as Aboriginal and/or Torres Strait Islander (p < 0.001). These differences are in line with the well-known observation that people from more socioeconomically advantaged backgrounds are more likely to be actively involved in research, who in turn are more likely to be older when they have their first child [13]. In addition, those with only one child are more likely to feel they have the time to be involved in research. These differences are very small, as, given the large sample size, even very small differences are statistically significant. A distinction needs to be made between statistical significance and practical or clinical significance. Researchers will need to make their own judgements regarding the impact of these differences on their specific research questions.

ORIGINS is fortunate to have access to additional routine data (from the Midwives Notification System) collected antenatally from some birthing mothers not enrolled at the Joondalup Health Centre (n = 4,131) during the study period. This allows a comparison of those enrolled in ORIGINS to those not enrolled in ORIGINS. These data are available to ORIGINS as part of the sub-project, “A Family Journey at JHC: analyses of routinely collected data”, which utilises waiver of consent that allows all routinely collected data at JHC to be available for analyses by ORIGINS. This allows a comparison of those enrolled in ORIGINS (n = 9,360 active and non-active mothers) to those not enrolled in ORIGIN (n = 4,131 mothers). Table 3 provides such a comparison on key demographic variables. Such a comparison allows researchers to be aware of any biases within the sample given the population ORIGINS aims to represent (that is, primarily those giving birth at the Joondalup Health Campus from 2017 to 2023). Note also that Table 3 comprises different comparisons to Table 2 as different data were available for this comparison.

The table shows that compared to non-ORIGINS birthing patients from JHC, ORIGINS participants are, on average, approximately one year older, over 2kg heavier, have slightly higher BMI (by 0.54 points), and are slightly more advantaged (scoring 0.37 decile points higher on the IRSAD) (p < 0.001, Wilcoxon rank sum test). Typically, those of higher socioeconomic status weigh less, on average, than those of a lower socioeconomic status [14], which contrasts with findings here. However, weight tends to increase with age [14], which may account for the heavier weight of ORIGINS versus non-ORIGINS patients, despite their average higher socioeconomic status. Again, these differences are small, so may have minimal clinical or practical impact [15], but the large sample size allows such differences to be detected. Further information about the ORIGINS cohort can be found in Talati et al (manuscript in preparation).

As noted above, ORIGINS has access to all data routinely collected at antenatal appointments, and at birth, for both non-active and active participants. All women giving birth at JHC, and their non-birthing partners, complete the Joondalup Health Questionnaire (JHQ). ORIGINS is fortunate to have access to all of these data during the study period, for both active and non-active participants, as well as for those not enrolled in ORIGINS. Hence, analyses can be performed to compare ORIGINS to non-ORIGINS patients, as well as non-active versus active ORIGINS participants with available JHC data.

The maternal digital JHQ collects information on participant characteristics and demographics, birth and pregnancy details, and information about fertility, mental health conditions and allergies. The paternal digital JHQ is similar but instead of pregnancy information, it collects information about marital status, smoking, drinking, drug-use and sexually transmitted diseases. These can be found in the Supplementary Appendix.

Other medical data collected via JHC are also shared with ORIGINS; that is, medical data collected antenatally via the Genie desktop digital patient data collection system at JHC, birthing and postnatal data collected via the Meditech hospital computer system, and data from the Midwives Notification System (MNS), a statutory data collection from the Manager of Administration and Health Information that collects data at birth. The MNS form undergoes frequent changes so the questions change; details of what questions are asked when can be found at Midwives Notification System on the Western Australian Department of Health website: (health.wa.gov.au). Participants also give ORIGINS permission to access their physical hospital files to check and supplement digital hospital data. Again, ORIGINS is fortunate to currently have access to MNS data for all mothers birthing at JHC during the study period, so comparisons can be made between ORIGINS and non-ORIGINS patients, as well as between non-active and active participants using available MNS data.

ORIGINS also has access to ultrasounds of ORIGINS participants (both non-active and active) during their pregnancy from the Perth Radiological Clinic. A legal and data transfer agreement was signed by both Perth Radiological Clinic and Telethon Kids Institute outlining the details of this agreement. Participant MRNs are used to match their data across both sites.

Similarly, ORIGINS has access to some limited data from participants’ routine blood tests during pregnancy (mother) and childbirth (mother and baby) through Western Diagnostics (a pathology provider that has partnered with ORIGINS). This allows ORIGINS access to the mother’s ferritin, haemoglobin levels, white blood cell count, platelets, HbA1C, Vitamin D, Glucose Tolerance Test (GTT), Thyroid-stimulating hormone (TSH) test, Group B Streptococcus (GBS) swab results (positive or negative) and BGN blood gases, and to the newborn infants’ full blood count, white blood cell count, platelets, serum bilirubin, vitamin D, PGL (to detect genes associated with hereditary paraganglioma or pheochromocytoma), c-reactive protein, cord Ph, lactate (BGN), and ferritin if available. The data in the pdf files is parsed using R-coding into csv format, but this is imperfect, and so needs to be manually checked and corrected by The ORIGINS Team.

All data collected at clinical appointments is stored in ORIGINS’ instance of the Research Electronic Data Capture (REDCap) database, such as anthropometry, morphology, dental, allergies (e.g. food, asthma, eczema, skin-prick test results), Nevisense [16] (uses electrical impedance to detect histological changes in the skin), ear otoscopy, optometry, joints, burns, medication use (including antibiotics), blood tests (e.g. serum ferritin levels, haemoglobin, HBA1c, and c-reactive protein), hospital admissions and referrals, family history of mental health issues, sleep and the cardiovascular, respiratory and neurological systems. Anthropometric measures resulting from the Bod Pod and the Paediatric Enclosed Air (PEA) Pod are also recorded. These use new technology that measures air displacement via plethysmography to determine body composition for children or adults and infants, respectively. These are now the gold standard for safe, repeatable and non-invasive assessment of body composition [17]. In addition, the novel Veggie Meter was used to detect skin carotenoids using refraction spectrometry as a proxy for recent fruit and vegetable intake [18]; this can be used to validate self-report dietary intake data. Finally, during clinical appointments, patients provide their immunisation records and their maternal and child health care ‘Purple Book’ (a logbook used by midwives and maternal child health nurses to track a newborn’s development and vaccinations into early childhood) to be photocopied for later manual entry into ORIGINS’ databases.

All of this data were collected at the 1, 3 and 5-year clinical appointments. These are shown in Table 5 below. Much of this data was also self-reported online via the online ORIGINS Core Questionnaire at other timepoints, so will appear in both Tables 5 and 7, but at different time-points.

Key administrative data are also collected, such as participant type (mother, non-birthing partner, child), participant status (withdrawn or lost-to-follow-up), consent level (active or non-active), contact details, sub-projects, and status of some measures and biosamples collected. Biosamples are tracked and managed via the online platform Open Specimen. A detailed overview is provided in D’Vaz et al [19] but a summary of collected biosamples is provided in Table 6.

ORIGINS online survey data primarily comprises the Core Questionnaire, administered at several timepoints throughout the ORIGINS journey. The survey comprises several items and measures administered at several different timepoints; not all measures were administered at all timepoints. What is administered when, and to whom, is summarised in Tables 7 and 8. Note that these tables only include measures collected at more than one timepoint; a full list of all measures is provided in Tables 1 to 5 in the Supplementary Appendix, for each domain. Incentives for Core Questionnaire completion are provided throughout the study period, primarily $20 gift cards for completion of the initial antenatal questionnaire and the completion of the 5-year Core Questionnaire, as well as the chance to go into a monthly draw for a $100 gift card.

Due to licensing arrangements, there are also three additional standardised measures collected by others on behalf of ORIGINS, the Conner’s Early Childhood Questionnaire, the Strengths and Difficulties Questionnaire and the Australia Eating Survey [20, 21]. The child version of the Australian Eating Survey was originally administered at 1 and 3-years postnatally. However, low completion rates, and issues with the birthing mother incorrectly entering her own data instead of the child’s, led the child version to be replaced by the 68-item Food Frequency Questionnaire [22]. This update resulted in greatly improved completion rates, from 20% to 58% completion per month.

ORIGINS has incorporated 51 sub-projects to date, many of which have collected their own data over different timepoints, while others have analysed existing data collected by ORIGINS. Sub-projects cover nine major research areas: Allergy, Inflammation and Immunity; Brain and Behaviour; Growth and Development; Nutrition and Metabolism; Environment and Lifestyle; Mental Health and Wellbeing; Microbiome; Parenting; and Oral Health. Nested sub-projects are required to return additional data they collect to The ORIGINS Project, so these data can be made available to other researchers. Further details about each sub-project can be found at The ORIGINS Project: Sub-projects website, https://originsproject.telethonkids.org.au/about-the-origins-project/.

ORIGINS can add measures to REDCap on behalf of sub-projects collecting additional data. For sub-projects conducting interventions, contact details of interested ORIGINS participants are passed to the sub-projects to organise data collection directly from participants.

The main strength of the ORIGINS data and biobank are the extent of the data collected and the novelty of much of this data. Since 2017, ORIGINS has collected over 300,000 biological samples and millions of datapoints to create an integrated data and biobank platform for researchers worldwide. ORIGINS has tracked mothers, non-birthing partners and children longitudinally to enable examination of multiple influences on child development from pre-conception, conception and early childhood across the various domains examined. This allows insights into the complex interactions between multifaceted exposures in the total lived environment (the ‘exposome’), and biological influence at the microscale, measured via multiomic analysis, including genomic, metabolomic and microbiomic techniques. Such analyses have not been possible in previous birth cohorts [11]. A further strength is that the extensive data collected by ORIGINS is supplemented by regularly merging with routine antenatal care data, as well as various other sources. In addition, as noted above, ORIGINS has utilised some recent and relatively novel data collection technologies not utilised by other birth cohorts, such as the BOD and PEA Pods, Nevisense [16], and the Veggie Meter. This allows novel insights into contemporary development.

ORIGINS is also unique in that it permits nesting of sub-projects to maximise the benefits of harmonised recruitment, observation and measurement. Moreover, ORIGINS allows sub-projects to test interventions, rather than just observing outcomes over time. Hence, ORIGINS provides a cost-effective way for researchers to conduct high-quality longitudinal observational and interventional studies using data already collected by ORIGINS, as well as adding their own. All data collected by ORIGINS sub-projects is returned to ORIGINS so other researchers can access these data. Further detail about this is provided below, when discussing how the ORIGINS data have been used.

A strength of the study is the timeliness of data availability. Often population datasets are several years old by the time they are cleaned, prepared and ready for public use [48]. By contrast, ORIGINS is able to clean and prepare data for use by researchers within months of its collection, so that the data are timely and relevant to researchers when they receive it.

In addition, ORIGINS’ different levels of consent provide a unique ability to compare those willing to actively participate in research (active participants) to those who are only willing to share their routinely collected data (non-active participants). ORIGINS also has ethics approval to access some data (JHC and MNS data) for antenatal hospital patients who did not enrol in ORIGINS. This allows ORIGINS to estimate bias inherent in their dataset by comparing these groups, such as that shown in Table 3.

Finally, ORIGINS has engaged community stakeholders throughout the development of the research platform in order to ensure the research design reflects community’s needs. Decision-making is shared with the community, and community feedback is utilised to check and then improve the research design. This has enhanced the design of the research in terms of acceptability and validity. In addition, ORIGINS has conducted several community events and has dedicated marketing personnel, so the project is well known to the community, particularly within the northern corridor of Perth where the Joondalup Health Campus is located. The ORIGINS Team actively follow up participants to reduce attrition, as well as ensuring the study provides tangible benefits to participants, such as extra paediatric check-ups and referrals. All of these strategies encourage maximum participation and retention, as well as enhancing the validity and quality of the data collected [50].

While The ORIGINS Project comprises over 50% of the population it aims to represent (i.e. those giving birth at JHC), and so can be regarded as a population-level cohort, a large proportion of the population did not consent to participate (47%). Younger and more affluent individuals are more likely to be included in digital data collections compared to older people, migrants, or those with a lower socio-economic status [51]. Although the recruitment of culturally and linguistically diverse (CALD) and Indigenous participants would have been beneficial, these particular groups weren’t prioritised. The recruitment material and questionnaires are in English and no translation services are offered, therefore participants must be sufficiently fluent in English to be able to participate. ORIGINS is in a unique position of accessing some data for those not enrolled in the study, so can do some comparisons of enrolled versus non-enrolled participants. Regardless, the latter are not included in the active cohort who complete the ORIGINS’ Core Questionnaire.

An additional potential issue is that data collected from some active participants is not purely observational, as is usually the case for population-level data, due to the impact of interventions with this group. Rather, true observational data are only available for non-active participants. This was a deliberate decision on behalf of ORIGINS as it was felt that the time for observation, or ‘watch and wait’ is over, and the opportunities to intervene in these crucial early years should not be squandered [11, 52]. Again, antenatal observational data are available for non-active participants, and for active participants not yet exposed to any interventions. Moreover, once ORIGINS links to other datasets, further observational data are available for non-active participants.

As noted above, although ORIGINS attempts to complete missing data, there is inevitably missing data and attrition across the timepoints. Some strategies have been put in place to reduce this as much as possible, but this is unavoidable, and is a well-known issue in longitudinal studies [53]. As The ORIGINS Project itself does not analyse data, but rather, provides data to sub-projects, researchers need to examine the impact of the missing data on their particular research (for example, whether the data are missing completely at random or not).

Relatedly, as ORIGINS collects information from a range of sources, there will be different levels of completion of such data across these sources for each participant. A comparison has been provided of any differences on some key demographic characteristics between those who completed each standardised measure for completed timepoints (antenatal and birth) but not for non-completed timepoints, nor for non-ORIGINS data sources. However, researchers accessing merged datasets provided by ORIGINS can conduct such comparisons.

Another potential issue is that ORIGINS integrates data that are routinely collected at JHC by health professionals for administrative purposes and for the patient’s health records. Hence, the primary aim of such data collection is not necessarily for research purposes, so the collected data may not meet the same quality criteria as data intentionally collected by ORIGINS, as discussed above.

ORIGINS aims to link with external State and Commonwealth datasets, such as Births, Deaths and Marriages, Medicare, NAPLAN, the Australian Early Development Census (AEDC), the WA Register of Developmental Abnormalities, and the Australian Immunisation Register (AIR). This will allow researchers to link antenatal and early childhood health and development information to educational and cognitive outcomes, to demographic outcomes (via Births, Deaths and Marriages) and to link the in-depth data of ORIGINS to other health outcomes, such as the impact of being vaccinated.

Data linkage will also allow greater comparison of non-active and active participants. Comparing non-active to active participants will allow researchers to ascertain the benefits of active participation in ORIGINS, such as extra paediatric check-ups and referrals, albeit, with the limitation that participants were not randomised to active versus non-active conditions.