Register      Login
Journal of Primary Health Care Journal of Primary Health Care Society
Journal of The Royal New Zealand College of General Practitioners
RESEARCH ARTICLE (Open Access)

How is enrolment with a general practice associated with subsequent use of the emergency department in Aotearoa New Zealand? A cohort study

Megan Pledger https://orcid.org/0000-0003-1669-8346 1 * , Maite Irurzun-Lopez 1 , Nisa Mohan 1 , Jacqueline Cumming 1
+ Author Affiliations
- Author Affiliations

1 Te Hikuwai Rangahau Hauora - Health Services Research Centre, Te Wāhanga Tātai Hauora – Wellington Faculty of Health, Te Herenga Waka - Victoria University of Wellington, PO Box 600, Wellington 6140, New Zealand.

* Correspondence to: megan.pledger@vuw.ac.nz

Handling Editor: Tim Stokes

Journal of Primary Health Care 16(2) 135-142 https://doi.org/10.1071/HC24023
Submitted: 19 February 2024  Accepted: 26 April 2024  Published: 9 May 2024

© 2024 The Author(s) (or their employer(s)). Published by CSIRO Publishing on behalf of The Royal New Zealand College of General Practitioners. This is an open access article distributed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND)

Abstract

Introduction

Around 5% of the people in Aotearoa New Zealand (NZ) are not enrolled with a general practice.

Aim

This study aimed to explore the utilisation of general practice by enrolment status and subsequent use of an emergency department.

Methods

We compared a cohort of respondents from New Zealand Health Surveys (2013/14–2018/19) on self-reported general practice utilisation and their substitutes, according to their enrolment status (enrolled and not enrolled). They were then followed up to examine their subsequent use of an emergency department. Time to an emergency department presentation was modelled with proportional hazards regression models with enrolment status as the explanatory variable. Confounding variables used were sex, age group, prioritised ethnicity, the New Zealand Deprivation Index and self-rated health.

Results

Those not enrolled were more likely to be young, male, Asian, more socioeconomically deprived and with better health status than those enrolled. Generally, those not enrolled utilised general practice services less. Those not enrolled who had used an emergency department were more likely to have used it as a substitute for general practice (40% vs 26%). Modelling showed that those not enrolled took longer to access an emergency department. Adjusting for confounding variables did not change that interpretation.

Discussion

Those not enrolled were younger and healthier and may have a perception that enrolment isn’t necessary. As a group, they were more likely to be socioeconomically deprived and to use an emergency department, which is free at a public hospital in NZ, as a substitute for primary care which suggests that cost may influence their choices.

Keywords: Aotearoa New Zealand, emergency department, general practice, health equity, primary health care, primary health care enrolment, primary healthcare organisations, survival analysis.

WHAT GAP THIS FILLS
What is already known: Around 5% of the population were not enrolled in a general practice in 2023. This has been found to differ by age, ethnicity and deprivation; with Māori and Pacific peoples (compared to non-Māori/non-Pacific people), young people and those who live in more socioeconomically deprived areas having lower enrolment rates.
What this study adds: Those not enrolled generally use fewer health services in general practice than the enrolled. They are less likely to use and take longer to access an emergency department compared to the enrolled but are more likely to use it for conditions that could be treated at a general practice.

Introduction

In many countries, primary health care (PHC) is the most common way to access a range of health care services. PHC, as well as providing health services in a community setting, can prevent or reduce the need for secondary and tertiary health care through health promotion and early medical intervention, and can facilitate access to other health and social care.1 Reducing the need for secondary and tertiary health care assists in lowering overall health system costs.2

In Aotearoa New Zealand (NZ), people are able to access PHC at a general practice. The extra step of enrolling gives people added benefits, which may include reduced user charges and access to specific health programmes, such as screening or counselling. However, enrolling may not always be possible at a desired general practice if the practice has closed its books to enrolment.3 In 2022, 33% of general practices in NZ had closed their books.4 There are also rural–urban disparities in general practice coverage and in socioeconomically deprived urban areas making access to enrolment difficult. In addition, certain classes of visitors to NZ and new migrants may not be eligible to enrol, at least initially.5 General practices benefit from enrolling patients as they gain funding for each patient enrolled, through a capitation formula.6

In 2023, around 5% of the population was not enrolled in a general practice and this has been found to differ by age, ethnicity and deprivation, with Māori and Pacific peoples (when compared explicitly to non-Māori/non-Pacific peoples), young people and those who live in more socioeconomically deprived areas having lower enrolment rates.7

The aim of this study was to better understand the way in which people who were not enrolled access care. On the one hand, those not enrolled may bypass general practice and access care through an emergency department (ED). On the other, those not enrolled may delay seeking care or face barriers to seeking care in general practice and end up in an ED when they become unwell.

For this study, we took a cohort of respondents from the New Zealand Health Surveys (NZHS) and ascertained who were enrolled in a general practice and who were not enrolled. From the survey data, we compared these groups by their use of health care in general practice and substitutes. Finally, we looked at their use of EDs in the years following the survey they participated in.

Methods

The data for this study were accessed via the Datalab at Statistics New Zealand.8 For a fuller account see the supplementary material accompanying this paper.

Five data sets were accessed and combined via the encrypted National Health Index. These were six NZHS surveys with data collected between 2013/14 and 2018/19 which provided self-reported information about the demographics, health care and health status of respondents. The Primary Health Organisation Enrolment data set and the National Enrolment Service data set gave information about enrolment status in general practice. The National Non-Admitted Patient Collection dataset gave information about presentations in EDs at public hospitals. The Mortality data set gave information about deaths.

Study period and follow-up

The study period for the collection of ED events started for each respondent at the end date of the survey they were in and finished on 30 June 2020, which meant the study period was between 1 and 6 years.

The statistical modelling looked at the time until an ED presentation. Follow-up started at the same point as the study period and lasted until either the respondent had an ED presentation, the respondent died or 30 June 2020; the last two events being censoring events. The timeline is available in Supplementary Fig. S1.

Statistical analysis

Data from the NZHS were analysed using proportions and means, and differences between enrolment groups were tested with two sample p-tests for differences in proportions and two sample t-tests for differences in means.

The length of time to the first ED presentation during the study period was modelled using survival analysis, specifically proportional hazards regression with enrolment status being the primary variable of interest (Model 1). The coefficients of this model were given in terms of hazard ratios which are a comparison of hazard rates between two groups. The greater the hazard rate for a group, the sooner a random respondent from that group will have an ED presentation.

The model was expanded twice: Model 2 with demographic variables – sex, age group, prioritised ethnicity and NZDep and Model 3 included a health variable – self-reported health.

The size of the dataset means that comparisons may be statistically significant when they are not practically significant. For example, the percentage of people going to a general practice in the last year for the two groups may be 76% and 78% which may be statistically significant at the 5% level; however, a difference between the two groups of 2% is not practically significant. When comparing two figures we comment when the difference is considered practically significant and statistically significant at the 5% level; otherwise, we refer to the figures as being similar. For percentages, this will be an absolute difference of 5%. For the differences in the average (1) number of visits or stays: 0.5 consults/visits, (2) consultation cost: $5, and (3) difference in time to first presentation: 30 days.

No ethics approval was sought for this project as this was a secondary analysis of administrative and survey data. Ethics approval had been sought previously by the Ministry of Health (MoH) from the New Zealand Health and Disability Multi-Region Ethics Committee for each survey.9

In the writing of this article, ChatGPT was occasionally used to suggest different ways to express ideas and explore word choices. This text was then further refined by the authors.10

Results

In the Datalab, the combined number of respondents across the six NZHS surveys was 72,243. Some respondents died before follow-up started (225, 0.3% of observations deleted). There were also some respondents in more than one survey for whom one random observation was chosen (516, 0.7%, observations deleted). This meant the analysis data set consisted of 71,502 respondents. See Supplementary Fig. S2 for a figurative description.

Demographics

Females (58%) were more likely to be enrolled than males (42%, see Table 1). Those enrolled were more likely to be middle-aged while those not enrolled were more likely to be younger.

Table 1.Enrolment groups by sociodemographic and health variables.

Enrolment groupsEnrolment groups
Not enrolledEnrolledNot enrolledEnrolled
N = 3681N = 67,821N = 3681N = 67,821
Stat95% CIStat95% CIStat95% CIStat95% CI
Sex (column %)Self-rated health (column %)
 Female37.7(36.1, 39.2)58.4(58.0, 58.7) Excellent18.1(16.8, 19.3)13.0(12.7, 13.2)
 Male62.3(60.8, 63.9)41.6(41.3, 42.0) Very good41.6(40.1, 43.2)38.8(38.5, 39.2)
Age group (column %) Good31.9(30.4, 33.5)34.2(33.8, 34.5)
 15–24 years22.2(20.8, 23.5)10.6(10.4, 10.8) Poor7.1(6.3, 7.9)11.0(10.8, 11.2)
 25–34 years29.0(27.5, 30.5)15.0(14.7, 15.3) Very poor1.2(0.9, 1.6)3.0(2.9, 3.2)
 35–44 years20.0(18.7, 21.3)16.3(16.0, 16.6)
 45–54 years14.5(13.4, 15.6)16.2(16.0, 16.5)Have you ever been told by a doctor that you have:
 55–64 years8.9(8.0, 9.8)16.6(16.3, 16.9) Had a heart attack (%)1.0(0.7, 1.3)4.4(4.2, 4.5)
 65–74 years3.7(3.1, 4.3)14.3(14.1, 14.6) Angina (%)1.0(0.7, 1.3)4.3(4.2, 4.5)
 75+ years1.8(1.4, 2.2)11.0(10.7, 11.2) Heart Failure (%)0.5(0.3, 0.7)2.7(2.6, 2.8)
Prioritised ethnicity (In priority order, column %) Other heart disease (%)3.3(2.7, 3.8)9.0(8.8, 9.2)
 Māori21.6(20.3, 22.9)20.6(20.3, 20.9) Stroke A (%)0.5(0.3, 0.7)2.2(2.1, 2.4)
 Pacific people7.2(6.3, 8.0)5.1(5.0, 5.3) Diabetes B (%)1.5(1.1, 1.9)7.6(7.4, 7.8)
 Asian21.8(20.4, 23.1)7.3(7.1, 7.5) Asthma (%)13.3(12.2, 14.4)21.4(21.1, 21.7)
 NZ European/Other49.5(47.9, 51.1)66.9(66.6, 67.3) Arthritis C (%)5.4(4.7, 6.1)22.0(21.7, 22.3)
NZ deprivation quintiles (column %) Depression D (%)6.9(6.1, 7.7)19.3(19.0, 19.6)
 1 (least deprived)10.5(9.5, 11.5)14.0(13.7, 14.2) Bipolar disorder D (%)0.4(0.2, 0.6)1.4(1.3, 1.5)
 213.9(12.7, 15.0)16.9(16.6, 17.2) Anxiety disorder D, E (%)4.2(3.5, 4.8)11.7(11.5, 12.0)
 320.4(19.1, 21.7)19.8(19.5, 20.1)
 424.7(23.3, 26.1)22.7(22.4, 23.0)
 5 (most deprived)30.6(29.1, 32.1)26.6(26.2, 26.9)
A Does not include transient ischaemic attacks.
B Does not include diabetes during pregnancy.
C Includes gout, lupus and psoriatic arthritis.
D Conditions that lasted or expected to last more than 6 months.
E Includes panic attacks, post-traumatic stress disorder (PTSD), phobias and obsessive-compulsive disorders.

Māori were almost equally represented in both groups (enrolled 22%, not enrolled 21%), with NZ European/Other more likely to be represented in the enrolled group (67% vs 49%) and the opposite for Asian peoples (7% vs 22%). Overall, respondents in the not-enrolled group were more likely to be in the three most socioeconomically deprived quintiles (76%) compared to the enrolled group (69%).

Table 1 also shows self-rated health and chronic conditions (any diagnosis in lifetime) by enrolment status. Those who were not enrolled were more likely to report excellent or very good health (60% vs 52%). Across the chronic conditions listed, those not enrolled were less likely to have a chronic condition than those enrolled. The biggest differences between the groups were for arthritis, depression and asthma.

General practice health care utilisation and substitutions

NZHS respondents were asked about their health services use. Only a third of the not-enrolled group had a consultation with a general practitioner (GP) in the year prior to their survey interview compared to 84% of the enrolled group (see Table 2). Those not enrolled who had seen a GP in the previous year had, on average, fewer consults (not enrolled 2.8, enrolled 4.0). The average cost of visiting a GP for a not-enrolled person ($42.90) was just over $10 more per visit than for someone enrolled. If they had not gone to see a GP when they needed to in the year prior, respondents were asked if it was because of cost; those not enrolled reported less unmet need because of cost which was statistically significantly different, but not practically different to those enrolled (14% vs 16%).

Table 2.Utilisation of health care in general practice and substitutions in the previous year by enrolment status.

Enrolment groupsSignificance E
Not-enrolledEnrolled
NStat95% CINStat95% CI
Had a GP consult (%) A367533.131.6–34.767,73784.183.9–84.4S, P
Number of GP consults (average, n) A, B12152.82.6–3.056,7993.93.95–4.03S, P
Cost of last GP consult (average, $NZ) B112842.940.3–45.454,12632.732.5–32.9S, P
Unmet need for GP consult because of cost (%) A367513.612.5–14.767,75816.215.9–16.5S
Had a nurse consult (%) A, C247511.610.4–12.943,87837.537.0–37.9S, P
Number of nurse consults (average, n) A, B, C2882.42.0–2.816,3802.42.4–2.5
Cost of last nurse consult (average, $NZ) B, C25812.79.3–16.014,73011.210.9–11.5
Had an after-hours consult A36817.06.2–7.867,78510.710.4–10.9S
Number of after-hours consults (average, n) A, B2581.41.3–1.672391.51.4–1.5
Cost of last after-hours consult (average, $NZ) B19863.156.9–69.4619256.255.3–57.1S, P
Unmet need for after-hours consult because of cost (%) A, D177613.011.4–14.634,14615.615.2–16.0S
Had an ED visit (%) A36819.68.7–10.667,78517.216.9–17.5S, P
Number of ED visits (average, n) A, B3541.51.3–1.611,6731.711.66–1.75S
Last visit to the ED for a condition that could have been treated at a medical centre (%) B34540.034.8–45.211,44226.225.4–27.0S, P
A Respondents were asked this question in terms of the previous year.
B Only asked of people who had the respective consult in the previous year.
C A nurse consult without a GP consult.
D For those needing an after-hours consult.
E S = statistical significance (P < 0.05), P = practical significance; P thresholds: 0.5 consults/visits, cost difference $5, percentage difference 5%.

Those enrolled were more likely to have had a nurse consult, but the number of consults and the cost of the last consult were similar between groups.

After-hours consults and ED visits are often used when there are barriers to accessing general practice, eg unable to get a timely appointment or because of the cost of a consultation. A similar proportion of each group reported having an after-hours consult with a similar number of consultations (see Table 2). However, the cost of an after-hours consult was $7 more for the not-enrolled group, on average ($63.10).

Those enrolled were more likely to report use of an ED in the year prior to their survey interview but those not enrolled were more likely to say that they used an ED for a condition that could have been treated at a medical centre (40% vs 26% out of those attending an ED).

Subsequent ED presentations

Table 3 describes the ED events that happened to the cohort in the study period, by enrolment status. Those not enrolled were less likely to have an ED presentation (26% vs 38%) during the study period, fewer presentations (2.2 vs 2.7) and had a longer time to their first presentation (585 days vs 546 days).

Table 3.Characteristics of ED presentations during the study period.

Enrolment groups
Not-enrolledEnrolled
NStat95% CINStat95% CISig C
ED presentation
 Had an ED presentation (%) A368126.224.7–27.667,82138.438.0–38.7S, P
 No of presentations (average, n) A, B9632.22.0–2.326,0312.72.7–2.8S, P
 Time to first presentation (average, days) B963585554–61626,031546541–552S, P
A Events happening during the study period.
B For those having an ED presentation during the study period.
C S = statistical significance (P < 0.05), P = practical significance; thresholds for practical significance: difference presentations: 0.5, difference in time to first presentation: 30 days, percentage difference: 5%.

Modelling of time to an ED presentation

The first model for time to an ED presentation, the univariate model, gives the hazard ratio for the not-enrolled group as 0.60. The hazard ratio being less than one means the time until an ED presentation is, on average, shorter for the enrolled group compared to the not-enrolled group (Table 4). Under the assumption of proportional hazards, a hazard rate of 0.60 means that a random person in the not-enrolled group has a probability of 37% of having an ED presentation before a random person in the enrolled group.

Table 4.Hazard ratios for Models 1, 2 and 3 for length of time until the first ED presentation during follow-up.

Model 1Model 2Model 3
Hazard ratio95% CIHazard ratio95% CIHazard ratio95% CI
Enrolment groups
 Not-enrolled0.60(0.56, 0.64)0.67(0.62, 0.71)0.70(0.65, 0.74)
 Enrolled111
Sex
 Female0.97(0.95, 0.99)0.98(0.96, 1.01)
 Male11
Age group
 15–24 years1.34(1.27, 1.40)1.38(1.31, 1.45)
 25–34 years1.18(1.13, 1.24)1.20(1.14, 1.25)
 35–44 years11
 45–54 years1.11(1.06, 1.16)1.08(1.04, 1.14)
 55–64 years1.23(1.17, 1.28)1.22(1.16, 1.27)
 65–74 years1.61(1.53, 1.68)1.63(1.56, 1.71)
 75+ years2.82(2.70, 2.95)2.83(2.71, 2.96)
Prioritised ethnicity
 Māori1.84(1.74, 1.95)1.77(1.67, 1.87)
 Pacific people1.66(1.54, 1.79)1.62(1.50, 1.74)
 Asian11
 NZ European/Other1.42(1.34, 1.50)1.43(1.35, 1.51)
NZ dep quintiles
 1 (least deprived)11
 21.12(1.07, 1.18)1.10(1.05, 1.16)
 31.27(1.22, 1.33)1.23(1.17, 1.29)
 41.40(1.34, 1.46)1.33(1.27, 1.39)
 5 (most deprived)1.62(1.55, 1.69)1.49(1.42, 1.55)
Self-rated health
 Excellent1
 Very good1.12(1.07, 1.17)
 Good1.43(1.37, 1.49)
 Poor1.98(1.89, 2.08)
 Very poor2.81(2.63, 3.00)

Note: for all variables, differences with the reference level have P values that are P < 0.0001 except for the difference between males and females for ED presentations (P = 0.0100 in Model 2 and P = 0.1450 in Model 3).

In Model 2, with the addition of confounders sex, age group, ethnicity and NZDep, the hazard ratio for not being enrolled increased to 0.67. A confounder analysis shows that the difference in hazard rates between Model 1 and Model 2 is almost solely due to the age group variable ie the enrolled group being younger explains some of the delay in having an ED presentation. In Model 3 (Table 4), the hazard ratio for not being enrolled is 0.70, indicating that the addition to the model of the variable self-rated health has only a minor impact on the hazard ratio for the time to an ED presentation for those not enrolled.

Discussion

Those not enrolled are more likely to be male, young and Asian compared to the enrolled. They are more likely to be living in the more socioeconomically deprived quintiles, to have better health and be less likely to have chronic diseases compared to those enrolled. Being younger, male and healthier suggests a perception of good health which explains why people choose not to enrol.

The young are also more likely to be transient, moving for education, work and new relationships which may hinder new enrolment or enrolment continuity. Those that are not enrolled are more likely to be young, for whom incomes tend to be lower, and live in more socioeconomically deprived quintiles, which suggests that affordability of accessing care may be an issue. While not-enrolled people consult GPs less frequently than the enrolled, they still experience unmet health care needs, particularly in terms of cost for both GP and after-hours consultations. Additionally, those not enrolled are more likely to opt for ED visits, where treatment is free, for conditions that could have been addressed in general practice.

While those not enrolled are more likely to be male, young, Asian, more socioeconomically deprived and in better health, they were not uniformly so. That there are respondents with chronic conditions eg diabetes and depression, who are not enrolled, indicates that they are a heterogeneous group with a range of needs.

Asian people in this cohort are the ethnic group least likely to be enrolled. This seems at odds with other research which found that Māori were the ethnic group least likely to be enrolled (compared to Pacific peoples and non-Māori/non-Pacific peoples).7 Starting in 2019, the MoH released a data series that allowed analysis of Asian people as their own separate ethnic group. The data from the June 2019 quarter gives the closest data available in time to the data reported here. It showed that enrolment rates per 100 population were 76% for Asian people compared to 82% for Māori.11,12 This supports our finding that the Asian population in the survey data had lower enrolment rates.

One hypothesis could be that the not-enrolled group could be disproportionately capturing new migrants who may be ineligible to enrol.5 New migrants were removed from the data set and the hazard rate for the not-enrolled group in Model 3 changed by 1.5% suggesting that eligibility to enrol is not driving the difference between the enrolled and not-enrolled results.

Altogether, the statistics have shown that those who are not enrolled are clearly getting fewer health services and there is reason to believe they are getting fewer health services than they need. An analysis of a complete collection of deaths in NZ between 2008 and 2017 found that those who died from an amenable death were more likely to not be enrolled, after adjusting for age, sex, ethnicity and socio-economic deprivation.13

The strength of this analysis is that it is able to draw on a large pool of respondents with wide-ranging information on their health service use. However, the NZHS does not reach people in outlying islands, those in hospital level care and those without a place of residence. In the 2018 census, just over 3500 people were in the group labelled ‘people without shelter’.14 It is more likely that this group had greater unmet need for health care and lower access to enrolment than the general population. In addition, those refusing to participate in the NZHS may have different characteristics to those that do.

Around 8% of the respondents in the NZHS were not included in the Integrated Data Infrastructure (IDI) dataset because they could not be matched to a National Health Index (NHI) number. It is extremely unlikely that those in the sample frame would not have a NHI number (personal communication, MoH) so not being able to make a match suggests either errors in recorded personal information or non-matches in home address. A study found that transience (moving more than three times in 3 years) occurs in around 5.5% of the population and 4.3% were considered vulnerable transients (had at least one housing occurrence in a socioeconomically deprived area).15 It seems likely that the unmatched respondents were more likely to not be enrolled and more likely to be socioeconomically deprived than the not-enrolled group studied here.

A limitation of this analysis is not knowing the interview dates of participants in the NZHS. To address this, we opted to start our study period at the conclusion of each survey period. As part of a sensitivity analysis, we re-analysed Model 3 but set the start date of follow-up to match the beginning of the survey period. This adjustment resulted in a 5% alteration of the hazard ratios for enrolment status, with the change being in the conservative direction.

There are significant differences between the enrolled and not-enrolled groups and the health services they use. To achieve the goals of equity in health and access to care,16 it is important to ensure that all populations enjoy equal opportunities to access health services, in particular PHC services, regardless of their economic means or other socioeconomic circumstances.

Supplementary material

Supplementary material is available online.

Data availability

The data are available in the datalab at Statistics New Zealand. Access requires a vetting process by Statistics New Zealand.

Conflicts of interest

The authors declare that they have no competing interests.

Declaration of funding

This work was funded by the Health Research Council grant – ‘Enhancing Primary Health Care Services to Improve Health in Aotearoa/New Zealand’ (Programme Grant 18/667) and the Lottery Health Research Funding Grant – ‘The challenge of closed books in primary care access, health outcomes and equity in NZ’ (LHR-2022-186638). The funders did not have any role in the design, collection, analysis, interpretation of data or writing and submission of the manuscript.

Disclaimer

Statement by Statistics New Zealand: Access to the data used in this study was provided by Stats NZ under conditions designed to give effect to the security and confidentiality provisions of the Statistics Act 1975. The results presented in this study are the work of the author, not Stats NZ or individual data suppliers. These results are not official statistics. They have been created for research purposes from the Integrated Data Infrastructure (IDI) which is carefully managed by Stats NZ. For more information about the IDI please visit https://www.stats.govt.nz/integrated-data/.

Acknowledgements

Thanks to the respondents of the New Zealand Health Survey, 2013/14–2018/19 for their participation in the surveys.

Author contributions

JC, MI.L, MP: conceptualisation, funding acquisition. MP: methodology, software, data curation, formal analysis. MP, MIL: writing – original draft. MIL, MP, JC, NM: writing – review and editing. MIL, MP, NM: project administration.

References

Te Whatu Ora Health New Zealand. How the health system works. 2020. Available athttps://www.wcdhb.health.nz/your-health/how-the-health-system-works/ [cited 28 February 2023].

Organization for Economic Co-operation and Development (OECD). Realising the Potential of Primary Health Care. Paris, France: OECD Publishing; 2020.

Irurzun-Lopez M, Pledger M, Mohan N, et al. ‘Closed Books’: restrictions to primary health care access in Aotearoa New Zealand – reporting results from a survey across general practices. N Z Med J 2024; 137(1591): 11-29.
| Crossref | Google Scholar | PubMed |

Pledger M, Irurzun-Lopez M, Mohan N, et al. An area-based description of closed book general practices in Aotearoa - New Zealand. J Prim Health Care 2023; 15(2): 128-34.
| Crossref | Google Scholar | PubMed |

Te Whatu Ora Health New Zealand. Eligibility checklists and decision trees. 2022. Available at https://www.tewhatuora.govt.nz/assets/Uploads/eligibility-checklist-oct-2015.pdf [cited 27 January 2023].

Cumming J. Aotearoa New Zealand’s Primary Health Care Strategy: Equity Enhancing in Policy and Practice? Lancet Global Health Commission on Financing Primary Health Care; 2022. Working Paper No. 6.

Pledger M, Mohan N, Silwal P, et al. The enrolment gap and the COVID-19 pandemic: an exploration of routinely collected primary care enrolment data from 2016 to 2023 in Aotearoa New Zealand. J Prim Health Care 2023; 15(4): 316-23.
| Crossref | Google Scholar | PubMed |

Statistics New Zealand Tatauranga Aotearoa. Integrated Data Infrastructure. 2022. Available at https://www.stats.govt.nz/integrated-data/integrated-data-infrastructure/ [cited 23 January 2023].

Manatū Hauora Minister of Health. Methodology Report: New Zealand Health Survey (2008-). 2024. Available at https://mohlibrary.softlinkhosting.co.nz/liberty/opac/search.do?mode=BASIC&openDetail=true&action=search&queryTerm=uuid%3D%225e0534400a5a01e25cbac58f0002db59%22&operator=OR&_open=1 [cited 29 January 2024].

10  OpenAI. ChatGPT 3.5 [Large language model]. 2024. Available at https://chat.openai.com/ [cited 19 February 2024].

11  Te Whatu Ora Health New Zealand. Enrolment with a general practice and primary health organisation (Access to primary care). 2023. Available at https://www.health.govt.nz/our-work/primary-health-care/enrolment-general-practice-and-primary-health-organisation [cited 15 August 2023].

12  Te Whatu Ora Health New Zealand. Populations web tool. 8 February 2023. Available at https://tewhatuora.shinyapps.io/populations-web-tool/ [cited 15 August 2023].

13  Silwal P, Lopez MI, Pledger M, et al. Association between enrolment with a Primary Health Care provider and amenable mortality: a national population-based analysis in Aotearoa New Zealand. PLoS One 2023; 18(2): e0281163.
| Crossref | Google Scholar | PubMed |

14  Statistics New Zealand Tatauranga Aotearoa. The state of housing in Aotearoa New Zealand. 15 December 2020. Available at https://www.stats.govt.nz/infographics/the-state-of-housing-in-aotearoa-new-zealand/ [cited 18 April 2024].

15  Jiang N, Pacheco G, Dasgupta K. Residential movement within New Zealand: Quantifying and characterising the transient population. Wellington, New Zealand: New Zealand Work Research Institute, Auckland University of Technology for Superu; 2017.

16  Manatū Hauora Ministry of Health. Health and disability system reforms. 2022. Available at https://www.health.govt.nz/new-zealand-health-system/overview-health-system/health-and-disability-system-reforms [cited 27 January 2023].