The impact of painful musculoskeletal conditions on management and outcomes of cancer: a linked electronic health record study

Date of Approval: 
2020-12-02 00:00:00
Lay Summary: 
Many people with cancer will already live with musculoskeletal pain (for example backache, knee pain). This pain, and associated poor function and sleep interference, may not be a health professional's priority but might affect the possibility of hospital admission when needed, or reduce effectiveness or receipt of appropriate treatment. This means people with musculoskeletal pain may spend longer in hospital, and have increased chance of worse outcomes from other illnesses. The objective of this study is to determine whether having musculoskeletal pain is associated with worse long term outcomes in people with cancer including increased likelihood of hospitalisation, length of time in hospital, and earlier mortality. We will use information from the Clinical Practice Research Datalink, linked to hospital data. We will analyse data of patients aged 45 years and over newly diagnosed with cancer. This condition is life-threatening, affects quality of life, is a common reason for going to hospital, and is of high NHS priority. We will determine whether these patients have previously consulted a doctor for musculoskeletal pain. We will compare likelihood of hospital admission, length of hospital stay, and risk of worse outcomes such as death and moving to palliative care between those with pre-existing musculoskeletal pain and those without, taking into account other illnesses. We will examine if findings vary by pain type such as back pain or knee pain, severity of pain, or by age.
Technical Summary: 
Cancer is life-threatening, affects quality of life, is a common reason for going to hospital, and is of high NHS priority. In people with long-term conditions, such as cancer, musculoskeletal pain is common but often neglected. Musculoskeletal comorbidity may impact on outcomes if pain (and associated restricted functioning and sleep interference) prevents or delays delivery of appropriate treatment or reduces its effectiveness. It may limit patients’ ability to manage other conditions at home, increasing likelihood of hospitalisation, extending time to discharge, and worsening outcomes. We aim to understand the extent of association between pre-existing musculoskeletal conditions and outcomes for patients with cancer, including time to hospital admission (for any reason) and long-term outcomes. We will analyse data of patients newly diagnosed with cancer and compare patients with a prior painful musculoskeletal condition requiring health care to patients without such a condition on rates of hospital admission, length of hospital stay, 30-day readmission, mortality, need for palliative care, and resource use and costs. Painful musculoskeletal conditions will be identified from primary care records in the 24-months prior to cancer diagnosis. We will compare hospital admission and readmission rates and long-term outcomes between those with and those without musculoskeletal comorbidity. Flexible parametric survival models will be used to model time to hospitalisation, mortality, and palliative care. Poisson regression will be used to determine differences in length of stay in hospital. We will include in further models proxies for pain severity (musculoskeletal referral, analgesia prescription). We will account for clustering by practice. Our findings will allow assessment of the potential for existing evidence-based management of musculoskeletal pain and associated disability to be targeted in these patients to make a substantial impact on outcomes of cancer diagnoses.
Health Outcomes to be Measured: 
1: Admission to hospital (for cancer and for any reason); 2: Length of stay in hospital; 3. Readmission to hospital within 30 days of discharge (for same or for different reason as initial admission); 4. Time to mortality; 5: Time to palliative care; 6: Provision of chemotherapy and radiotherapy, prescription of hormone modulation treatment for breast cancer and anti-androgen hormone therapy for prostate cancer; 7. Cumulative health care use and costs over 5 years after index date. Primary care data will include number and type of consultation with each health care professional, prescriptions, tests and investigations. Secondary care utilisation includes referral, type of admission, length of stay, diagnosis, and procedures undertaken.
Application Number: 

Kelvin Jordan - Chief Investigator - Keele University
Kelvin Jordan - Corresponding Applicant - Keele University
Alyson Huntley - Collaborator - University of Bristol
Christian Mallen - Collaborator - Keele University
Felix Achana - Collaborator - University of Oxford
James Bailey - Collaborator - Keele University
John Edwards - Collaborator - Keele University
Kayleigh Mason - Collaborator - Keele University
Mamas Mamas - Collaborator - Keele University
Martin Frisher - Collaborator - Keele University
Simon White - Collaborator - Keele University
stephen tatton - Collaborator - Keele University

HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation