Health of mothers of children with a life-limiting condition: a protocol for comparative cohort study using the Clinical Practice Research Datalink

Introduction There are now nearly 50 000 children with a life-limiting or life-threatening condition in the UK. These include conditions where there is no reasonable hope of cure and from which they will die, as well as conditions for which curative treatment may be feasible but can fail, for example, cancer or heart failure. Having a child with a life-limiting condition involves being a coordinator and provider of healthcare in addition to the responsibilities and pressures of parenting a child who is expected to die young. This adversely affects the health and well-being of these mothers and affects their ability to care for their child, but the extent of the impact is poorly understood. This study aims to quantify the incidence and nature of mental and physical morbidity in mothers of children with a life-limiting condition, their healthcare use and to assess whether there is a relationship between the health of the mother and the child’s condition. Methods and analysis A comparative cohort study using data from the Clinical Practice Research Datalink and linked hospital data will include three groups of children and their mothers (those with a life-limiting condition, those with a chronic condition and those with no long-term health condition total=20 000 mother–child dyads). Incidence rates and incidence rate ratios will be used to quantify and compare the outcomes between groups with multivariable regression modelling used to assess the relationship between the child’s disease trajectory and mother’s health. Ethics and dissemination This study protocol has approval from the Independent Scientific Advisory Committee for the UK Medicines and Healthcare products Regulatory Agency Database Research. The results of this study will be reported according to the STROBE and RECORD guidelines. There will also be a lay summary for parents which will be available to download from the Martin House Research Centre website (www.york.ac.uk/mhrc).

I, the Submitting Author has the right to grant and does grant on behalf of all authors of the Work (as defined in the below author licence), an exclusive licence and/or a non-exclusive licence for contributions from authors who are: i) UK Crown employees; ii) where BMJ has agreed a CC-BY licence shall apply, and/or iii) in accordance with the terms applicable for US Federal Government officers or employees acting as part of their official duties; on a worldwide, perpetual, irrevocable, royalty-free basis to BMJ Publishing Group Ltd ("BMJ") its licensees and where the relevant Journal is co-owned by BMJ to the co-owners of the Journal, to publish the Work in this journal and any other BMJ products and to exploit all rights, as set out in our licence.
The Submitting Author accepts and understands that any supply made under these terms is made by BMJ to the Submitting Author unless you are acting as an employee on behalf of your employer or a postgraduate student of an affiliated institution which is paying any applicable article publishing charge ("APC") for Open Access articles. Where the Submitting Author wishes to make the Work available on an Open Access basis (and intends to pay the relevant APC), the terms of reuse of such Open Access shall be governed by a Creative Commons licence -details of these licences and which Creative Commons licence will apply to this Work are set out in our licence referred to above.
Other than as permitted in any relevant BMJ Author's Self Archiving Policies, I confirm this Work has not been accepted for publication elsewhere, is not being considered for publication elsewhere and does not duplicate material already published. I confirm all authors consent to publication of this Work and authorise the granting of this licence. There are now nearly 50,000 children with a life-limiting or life-threatening conditions in the UK. These include conditions where there is no reasonable hope of cure and from which they will die, as well as conditions for which curative treatment may be feasible but can fail, e.g. cancer or heart failure. Having a child with a life-limiting condition involves being a coordinator and provider of healthcare in addition to the responsibilities and pressures of parenting a child who is expected to die young. This adversely affects the health and wellbeing of these mothers and affects their ability to care for their child, but the extent of the impact is poorly understood.
This study aims to quantify the incidence and nature of mental and physical morbidity in mothers of children with a life-limiting condition, their healthcare use and to assess whether there is a relationship between the health of the mother and the child's condition.

Methods and analysis
A comparative cohort study using data from the Clinical Practice Research Datalink and linked hospital data will include three groups of children and their mothers (those with a life-limiting condition, those with a chronic conditions and those with no long term health condition total =20,000 mother-child dyads). Incidence rates and incidence rate ratios will be used to quantify and compare the outcomes between groups with multivariable regression modelling used to assess the relationship between the child's disease trajectory and mother's health.

Ethics and dissemination
This study protocol has approval from the Independent Scientific Advisory Committee for the UK Medicines and Healthcare products Regulatory Agency Database Research. The results of this study will be reported according to the STROBE and RECORD guidelines. There will also be a lay summary for parents which will be available to download from the Martin House Research Centre website (www.york.ac.uk/mhrc ).  1 These include conditions for which there is no reasonable hope of cure and from which children or young people will die, as well as conditions for which curative treatment may be feasible but can fail, such as cancer or heart failure. In children and young people, more than 300 diagnoses are life-limiting or life-threatening 2 , including duchenne muscular dystrophy, severe cerebral palsy, neurodegenerative conditions, and severe congenital anomalies.
Although many of the individual diagnoses are rare, collectively life-limiting conditions affect more children and young people than more common single long term conditions, such as diabetes mellitus 3 . Many of these children are living longer due to the use of medical technologies, e.g. ventilation and gastrostomy feeding, and more aggressive treatment of complications and they are often high users of healthcare services 4 . The parents of these children have an added layer of 'pressure' from having a child with a chronic condition who also has a shortened life expectancy 5 . Having a child with a life-limiting condition often involves being a coordinator and provider of healthcare as well as being a parent, and this responsibility exists 24 hours a day, 7 days a week, especially for mothers. The health of these mothers is important, both for their own well-being, and also in terms of their ability to care for their child. Currently the NHS and statutory services for these children rarely extend to providing care for the parents. The focus is on mothers rather than fathers in this study as mothers are the main carer for the majority of these children.
The need for further research on the emotional and psychological support and interventions for parents or carers of children with a life-limiting condition was highlighted recently in NICE guidelines on end of life care in children and young people 6 . This stated that research was needed on 'What emotional support do children and young people with life-limiting conditions and their parents or carers need, and how would they like these needs to be addressed? The NICE guidelines 6 also noted that no studies had quantified the psychological/ mental health of mothers of children with life-limiting condition. There have been some attempts to quantify these in children with special needs 7 or specific disabilities 8 9 which have shown higher levels of parental distress or emotional problems than parents of healthy children (36% cf 20%). These however, do not address the specific needs of those with lifelimiting conditions or the added burden that parents of children with a life-limiting condition face, knowing their child is likely to die.
Evidence about the physical health of mothers with life-limiting condition is also sparse, although two studies in mothers of children with disabilities found higher prevalence of physical conditions compared to mothers of healthy children. For example, prevalence of back pain and hypertension was 35.2% and 24.7% respectively in mothers of children with disabilities, compared to 26.7% and 19.1% in mothers of healthy children. However, these studies were cross sectional, self-reported and were not explicitly for mothers of children with a life-limiting condition 9 10 .. To date, no study using a nationally representative sample has assessed the incidence or prevalence of mental and physical health conditions in mothers of children with a life-limiting condition. Quantifying the extent and nature of both physical and mental health conditions in these mothers is important when trying to develop and target appropriate preventative or treatment interventions for this population.
This study aims to quantify the incidence and nature of mental and physical morbidity in mothers of children with a life-limiting condition, their healthcare use and to assess whether there is a relationship between the health of the mother and the child's condition.

1.
What is the nature and incidence of mental and physical morbidity in mothers of children with a life-limiting condition? 2.
What is the relationship between the health of the mother and the child's condition?
3. How many primary care visits do these mothers have compared to other mothers?
4. How many hospital admissions do these mothers have compared to other mothers?

5.
What are the resource costs of healthcare for these mothers?

Data sources and sample selection
This will be an observational comparative cohort study design using data from the Clinical The index children will be extracted from the CPRD mother baby linkage, where the mother has at least 1 years registration and where they eligible for HES linkage (resident in England).
Children will be grouped in to those with a life-limiting condition (n~5000) or chronic condition that is not life-limiting e.g. diabetes, asthma (n~5000) and they will be identified within the CPRD dataset between 2007-2017, using READ codes in the CPRD data and ICD10 codes in the HES data 1 (see Table 1) . The development of these ICD-10 code list have been  15 . Subjects will be matched by year of birth, sex and region to the control group of healthy children (n~10,000) with no long term conditions.
All primary and the linked secondary care data will be extracted for each mother and child dyad.

Sample size
A sample size calculation based on 80% power, 5% significance to detect a 5% difference in prevalence in maternal anxiety/depression between the groups (from 15-20% prevalence) indicates a minimum of 1,131 participants required in each group. Given the heterogeneity in the life-limiting condition group, and to facilitate analyses within life-limiting condition subgroups, we will aim to have 5,000 participants in each of the life-limiting condition and chronic disease groups, with 10,000 in the control (no long-term condition) group. The total sample size will be 40,000 (20,000 children and 20,000 mothers).

Exposures
The key exposure is the disease status of the child (life-limiting condition/chronic disease (asthma or diabetes)/ no long-term condition). These groups will be identified using previously developed READ and ICD 10 diagnostic code lists (see supplementary material).
The use of both READ and ICD 10 codes will enable identification of relevant diagnoses from either the primary care records (READ) or hospital admission data (ICD 10).

Outcomes
The outcomes for the mothers will be recorded in either the primary care data or the linked secondary care datasets. The dates of these outcomes will be assessed in relation to the date of the child's diagnoses.

Incidence of mental health and physical health conditions
 Maternal mental health diagnoses, including both common mental illness and severe mental illness. These will be identified using diagnostic READ codes and relevant

Statistical methods
All analyses will be undertaken using STATA version 15 19 . All analyses will adjust for known clinical and demographic confounders of importance (age, co-morbidities, socioeconomic status, previous use of primary care services, etc.). The left and right censored nature of these data will be taken into account when undertaking these analyses.

Incidence of mental health and physical health conditions
Incidence rates of the physical and mental health conditions will be calculated in each group of mothers by dividing the number of cases in each group by the person-time at risk in each group. A comparison with the incidence of these conditions in the three groups of mothers will be undertaken using incidence rate ratios.

The association of maternal health with the child's disease trajectory
The presence of physical or mental health diagnoses in the mother will be the outcome variable of interest in a multivariable model with a key covariate being the presence of prolonged or repeated hospital admission for the child. A prolonged admission is one that lasts > 14 days and repeated admissions will be defined as more than 2 admissions in a 6 month time period.
Healthcare use and resources of the mother Using multivariable Poisson regression modelling for the count of GP visits, referrals and hospital admissions with the key exposure of disease status of the child (life-limiting condition /Chronic disease/ no long-term condition). Multilevel modelling will be used to account for repeated measures within individuals. For the analyses of healthcare costs, we will use information from Unit Costs of Health and Social Care 20 , the NHS national tariffs 21 11 We expect there to be missing data for some of the key covariates in the data e.g. ethnicity.
We will assess the level of missingness across all sources of these variable (HES and CPRD). If appropriate then we will use multiple imputation techniques 23 to address this in order to avoid the potential bias from only undertaking complete case analyses.

PATIENT AND PUBLIC INVOLVEMENT
The views of parents and carers of children with a life-limiting condition have informed the development of this study. Initial ideas about the topic area and planned study were discussed with a group of 10 parents of children with a life-limiting condition at Martin House Children's Hospice. They helped to refine the research question and future stages of this programme of work. The Family Advisory Board of the Martin House Research Centre (www.york.ac.uk/mhrc) will have involvement in this study with an emphasis on the preparation of the lay summaries and dissemination to parents and children.

ETHICS AND DISSEMINATION
This study protocol has approval from the Independent Scientific Advisory Committee for the UK Medicines and Healthcare products Regulatory Agency Database Research (protocol 18_313). They have NREC approval for release of pseudonymised data for observational research.

Introduction
There are now nearly 50,000 children with a life-limiting or life-threatening conditions in the UK. These include conditions where there is no reasonable hope of cure and from which they will die, as well as conditions for which curative treatment may be feasible but can fail, e.g.
cancer or heart failure. Having a child with a life-limiting condition involves being a coordinator and provider of healthcare in addition to the responsibilities and pressures of parenting a child who is expected to die young. This adversely affects the health and wellbeing of these mothers and affects their ability to care for their child, but the extent of the impact is poorly understood.
This study aims to quantify the incidence and nature of mental and physical morbidity in mothers of children with a life-limiting condition, their healthcare use and to assess whether there is a relationship between the health of the mother and the child's condition.

Methods and analysis
A comparative cohort study using data from the Clinical Practice Research Datalink (CPRD) and linked hospital data will include three groups of children and their mothers (those with a life-limiting condition, those with a chronic conditions and those with no long term health condition total =20,000 mother-child dyads). Incidence rates and incidence rate ratios will be used to quantify and compare the outcomes between groups with multivariable regression modelling used to assess the relationship between the child's disease trajectory and mother's health.

Ethics and dissemination
This study protocol has approval from the Independent Scientific Advisory Committee for the UK Medicines and Healthcare products Regulatory Agency Database Research. The results of this study will be reported according to the STROBE and RECORD guidelines. There will also

Strengths and Limitations of this study
This study is utilising a large, longitudinal, nationally representative primary care data source.
The Clinical Practice Research Datalink data enables the identification of mother and child dyads and linkage to secondary care data.
This study is reliant on identification of the cohort of interest by diagnostic coding therefore on the quality of diagnostic coding within the CPRD and HES datasets.

INTRODUCTION
The number of children with life-limiting or life-threatening conditions has been rising, with latest figures estimating 49,000 children and young people with a life-limiting condition in the UK. 1 These include conditions for which there is no reasonable hope of cure and from which children or young people will die, as well as conditions for which curative treatment may be feasible but can fail, such as cancer or heart failure. In children and young people, more than 300 diagnoses are life-limiting or life-threatening 2 , including duchenne muscular dystrophy, severe cerebral palsy, neurodegenerative conditions, and severe congenital anomalies.
Although many of the individual diagnoses are rare, collectively life-limiting conditions affect more children and young people than more common single long term conditions, such as diabetes mellitus 3 . Many of these children are living longer due to the use of medical technologies, e.g. ventilation and gastrostomy feeding, and more aggressive treatment of complications and they are often high users of healthcare services 4 . The parents of these children have an added layer of 'pressure' from having a child with a chronic condition who also has a shortened life expectancy 5 . Having a child with a life-limiting condition often involves being a coordinator and provider of healthcare as well as being a parent, and this responsibility exists 24 hours a day, 7 days a week, especially for mothers. The health of these mothers is important, both for their own well-being, and also in terms of their ability to care  This study aims to quantify the incidence and nature of mental and physical morbidity in mothers of children with a life-limiting condition, their healthcare use and to assess whether there is a relationship between the health of the mother and the child's condition.

1.
What is the nature and incidence of mental and physical morbidity in mothers of children with a life-limiting condition?

2.
What is the relationship between the health of the mother and the child's condition?
3. How many primary care visits do these mothers have compared to other mothers?

4.
How many hospital admissions do these mothers have compared to other mothers?

5.
What are the resource costs of healthcare for these mothers?

Data sources and sample selection
This will be an observational comparative cohort study design using data from the Clinical The index children will be extracted from the CPRD mother baby linkage, where the mother has at least 1 years registration and where they eligible for HES linkage (resident in England).
Children will be grouped in to those with a life-limiting condition (n~5000) or chronic condition that is not life-limiting e.g. diabetes, asthma (n~5000) and they will be identified within the CPRD dataset between 2007-2017, using READ codes in the CPRD data and ICD10 codes in the HES data 1 (see Table 1) . The development of these ICD-10 code list have been year of birth, sex and region to the control group of healthy children (n~10,000) with no long term conditions.

Sample size
A sample size calculation based on 80% power, 5% significance to detect a 5% difference in prevalence in maternal anxiety/depression between the groups (from 15-20% prevalence) indicates a minimum of 1,131 participants required in each group. Given the heterogeneity in the life-limiting condition group, and to facilitate analyses within life-limiting condition subgroups, we will aim to have 5,000 participants in each of the life-limiting condition and chronic disease groups, with 10,000 in the control (no long-term condition) group. The total sample size will be 40,000 (20,000 children and 20,000 mothers).

Exposures
The key exposure is the disease status of the child (life-limiting condition/chronic disease (asthma or diabetes)/ no long-term condition). These groups will be identified using previously developed READ and ICD 10 diagnostic code lists (see supplementary material).
The use of both READ and ICD 10 codes will enable identification of relevant diagnoses from either the primary care records (READ) or hospital admission data (ICD 10).

Outcomes
The outcomes for the mothers will be recorded in either the primary care data or the linked secondary care datasets. The dates of these outcomes will be assessed in relation to the date of the child's diagnoses.

Incidence of mental health and physical health conditions
 Maternal mental health diagnoses, including both common mental illness and severe mental illness. These will be identified using diagnostic READ codes and relevant  16 will be utilised. The linked MHMDS will indicate the presence of a more serious mental health conditions.  Maternal physical diagnoses, including e.g. obesity, hypertension, musculoskeletal problems, and cardiovascular disease. These will be identified using diagnostic READ codes, relevant prescription data, HES data, and related biometric data (including blood pressure, BMI, and total cholesterol).
The association of maternal health with the child's disease trajectory  The number of A&E attendances, hospital admissions and length of stay for the children (HES data).
Healthcare use and resources of the mother  Referrals to other services, especially secondary services for mothers (CPRD and HES data).
 Uptake of cervical screening by mothers (CPRD data).  The number of A&E attendances, hospital admissions and length of stay for mothers (HES data).

Confounding
The following clinical and demographic confounders will be considered in the analyses:  Ethnicity, this will be classified according to the census 2011 categories 17 using data from CPRD and HES data.

Statistical methods
All analyses will be undertaken using STATA version 15 19 . All analyses will adjust for known clinical and demographic confounders of importance (age, co-morbidities, socioeconomic status, previous use of primary care services, etc.). The left and right censored nature of these data will be taken into account when undertaking these analyses.

Incidence of mental health and physical health conditions
Incidence rates of the physical and mental health conditions will be calculated in each group of mothers by dividing the number of cases in each group by the person-time at risk in each group. Incident cases will be counted only after the diagnoses in the child. A comparison with the incidence of these conditions in the three groups of mothers will be undertaken using incidence rate ratios.

The association of maternal health with the child's disease trajectory
The presence of physical or mental health diagnoses in the mother will be the outcome variable of interest in a multivariable model with a key covariate being the presence of prolonged or repeated hospital admission for the child. A prolonged admission is one that lasts > 14 days and repeated admissions will be defined as more than 2 admissions in a 6 month time period.
Healthcare use and resources of the mother

Missing data
We expect there to be missing data for some of the key covariates in the data e.g. ethnicity.
We will assess the level of missingness across all sources of these variable (HES and CPRD). If appropriate then we will use multiple imputation techniques 23 to address this in order to avoid the potential bias from only undertaking complete case analyses.

PATIENT AND PUBLIC INVOLVEMENT
The views of parents and carers of children with a life-limiting condition have informed the development of this study. Initial ideas about the topic area and planned study were discussed with a group of 10 parents of children with a life-limiting condition at Martin House Children's Hospice. They helped to refine the research question and future stages of this programme of work. The Family Advisory Board of the Martin House Research Centre (www.york.ac.uk/mhrc) will have involvement in this study with an emphasis on the preparation of the lay summaries and dissemination to parents and children.

ETHICS AND DISSEMINATION
This study protocol has approval from the Independent Scientific Advisory Committee for the UK Medicines and Healthcare products Regulatory Agency Database Research (protocol 18_313). They have NREC approval for release of pseudonymised data for observational research.
This study will be complete by Feb 2021.
The results of this study will be reported according to the STROBE and RECORD guidelines (ref). There were also be a lay summary for parents which will be available to download from the Martin House Research Centre website (www.york.ac.uk/mhrc ).

Introduction
There are now nearly 50,000 children with a life-limiting or life-threatening conditions in the UK. These include conditions where there is no reasonable hope of cure and from which they will die, as well as conditions for which curative treatment may be feasible but can fail, e.g.
cancer or heart failure. Having a child with a life-limiting condition involves being a coordinator and provider of healthcare in addition to the responsibilities and pressures of parenting a child who is expected to die young. This adversely affects the health and wellbeing of these mothers and affects their ability to care for their child, but the extent of the impact is poorly understood.
This study aims to quantify the incidence and nature of mental and physical morbidity in mothers of children with a life-limiting condition, their healthcare use and to assess whether there is a relationship between the health of the mother and the child's condition.

Methods and analysis
A comparative cohort study using data from the Clinical Practice Research Datalink (CPRD) and linked hospital data will include three groups of children and their mothers (those with a life-limiting condition, those with a chronic conditions and those with no long term health condition total =20,000 mother-child dyads). Incidence rates and incidence rate ratios will be used to quantify and compare the outcomes between groups with multivariable regression modelling used to assess the relationship between the child's disease trajectory and mother's health.

Strengths and Limitations of this study
This study is utilising a large, longitudinal, nationally representative primary care data source.
The Clinical Practice Research Datalink data enables the identification of mother and child dyads and linkage to secondary care data.
This study is reliant on identification of the cohort of interest by diagnostic coding therefore on the quality of diagnostic coding within the CPRD and HES datasets.

INTRODUCTION
The number of children with life-limiting or life-threatening conditions has been rising, with latest figures estimating 49,000 children and young people with a life-limiting condition in the UK. 1 These include conditions for which there is no reasonable hope of cure and from which children or young people will die, as well as conditions for which curative treatment may be feasible but can fail, such as cancer or heart failure. In children and young people, more than 300 diagnoses are life-limiting or life-threatening 2 , including duchenne muscular dystrophy, severe cerebral palsy, neurodegenerative conditions, and severe congenital anomalies.
Although many of the individual diagnoses are rare, collectively life-limiting conditions affect more children and young people than more common single long term conditions, such as diabetes mellitus 3 . Many of these children are living longer due to the use of medical technologies, e.g. ventilation and gastrostomy feeding, and more aggressive treatment of complications and they are often high users of healthcare services 4 . The parents of these children have an added layer of 'pressure' from having a child with a chronic condition who also has a shortened life expectancy 5 . Having a child with a life-limiting condition often involves being a coordinator and provider of healthcare as well as being a parent, and this responsibility exists 24 hours a day, 7 days a week, especially for mothers. The health of these mothers is important, both for their own well-being, and also in terms of their ability to care  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   5 for their child. Currently the National Health Service and statutory services for these children rarely extend to providing care for the parents. The focus is on mothers rather than fathers in this study as mothers are the main carer for the majority of these children.
The This study aims to quantify the incidence and nature of mental and physical morbidity in mothers of children with a life-limiting condition, their healthcare use and to assess whether there is a relationship between the health of the mother and the child's condition. What is the relationship between the health of the mother and the child's condition?
3. How many primary care visits do these mothers have compared to other mothers?

4.
How many hospital admissions do these mothers have compared to other mothers?

5.
What are the resource costs of healthcare for these mothers?

Data sources and sample selection
This will be an observational comparative cohort study design using data from Statistics death certificate data. The CPRD provides a unique opportunity to assess the relationship between maternal morbidity and their child health due to the longitudinal nature of data collection (since 1987) and the ability to link mothers and children via their motherbaby link algorithm 14 . It is not currently possible to link babies to fathers in this dataset.
The index children will be extracted from the CPRD mother baby linkage, where the mother has at least 1 years registration and where they eligible for HES linkage (resident in England).
Children will be grouped in to those with a life-limiting condition (n~5000) or chronic condition that is not life-limiting e.g. diabetes, asthma (n~5000) and they will be identified within the CPRD dataset between 2007-2017, using READ codes in the CPRD data and ICD10 codes in the HES data 1 (see Table 1) . The development of these ICD-10 code list have been  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   7 described previously for life-limiting 1 and chronic conditions 15 . Subjects will be matched by year of birth, sex and region to the control group of healthy children (n~10,000) with no long term conditions.

Exposures
The key exposure is the disease status of the child (life-limiting condition/chronic disease (asthma or diabetes)/ no long-term condition). These groups will be identified using previously developed READ and ICD 10 diagnostic code lists (see supplementary material).
The use of both READ and ICD 10 codes will enable identification of relevant diagnoses from either the primary care records (READ) or hospital admission data (ICD 10).

Outcomes
The outcomes for the mothers will be recorded in either the primary care data or the linked secondary care datasets. The dates of these outcomes will be assessed in relation to the date of the child's diagnoses.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y 10 Incidence of mental health and physical health conditions  Maternal mental health diagnoses, including both common mental illness and severe mental illness. These will be identified using diagnostic READ codes and relevant prescription data. A previously developed READ code algorithm 16 will be utilised. The linked MHMDS will indicate the presence of a more serious mental health conditions.  Maternal physical diagnoses, including e.g. obesity, hypertension, musculoskeletal problems, and cardiovascular disease. These will be identified using diagnostic READ codes, relevant prescription data, HES data, and related biometric data (including blood pressure, BMI, and total cholesterol).

The association of maternal health with the child's disease trajectory
 The number of A&E attendances, hospital admissions and length of stay for the children (HES data).
Healthcare use and resources of the mother  Referrals to other services, especially secondary services for mothers (CPRD and HES data).
 Uptake of cervical screening by mothers (CPRD data).  The number of A&E attendances, hospital admissions and length of stay for mothers (HES data).

Confounding
The following clinical and demographic confounders will be considered in the analyses:  Ethnicity, this will be classified according to the census 2011 categories 17 using data from CPRD and HES data.

Statistical methods
All analyses will be undertaken using STATA version 15 19 . All analyses will adjust for known clinical and demographic confounders of importance (age, co-morbidities, socioeconomic status, previous use of primary care services, etc.). The left and right censored nature of these data will be taken into account when undertaking these analyses.

Incidence of mental health and physical health conditions
Incidence rates of the physical and mental health conditions will be calculated in each group of mothers by dividing the number of cases in each group by the person-time at risk in each group. Incident cases will be counted only after the diagnoses in the child. A comparison with the incidence of these conditions in the three groups of mothers will be undertaken using incidence rate ratios.

The association of maternal health with the child's disease trajectory
The presence of physical or mental health diagnoses in the mother will be the outcome variable of interest in a multivariable model with a key covariate being the presence of prolonged or repeated hospital admission for the child. A prolonged admission is one that lasts > 14 days and repeated admissions will be defined as more than 2 admissions in a 6 month time period.

Missing data
We expect there to be missing data for some of the key covariates in the data e.g. ethnicity.
We will assess the level of missingness across all sources of these variable (HES and CPRD). If appropriate then we will use multiple imputation techniques 23 to address this in order to avoid the potential bias from only undertaking complete case analyses.

PATIENT AND PUBLIC INVOLVEMENT
The views of parents and carers of children with a life-limiting condition have informed the development of this study. Initial ideas about the topic area and planned study were discussed with a group of 10 parents of children with a life-limiting condition at Martin House Children's Hospice. They helped to refine the research question and future stages of this programme of work. The Family Advisory Board of the Martin House Research Centre (www.york.ac.uk/mhrc) will have involvement in this study with an emphasis on the preparation of the lay summaries and dissemination to parents and children.

ETHICS AND DISSEMINATION
This study protocol has approval from the Independent Scientific Advisory Committee for the UK Medicines and Healthcare products Regulatory Agency Database Research (protocol 18_313). They have NREC approval for release of pseudonymised data for observational research.