Keywords
pain, ageing, overprescribing, big data, medication
-
Views
6
Downloads
How to cite this article
G. Spicer S, A Payne R, Schofield P et al. Helping to Optimise Pain control in the Elderly experiencing Analgesic Overprescribing (HOPE-AO): protocol for a descriptive epidemiological analysis of patterns of analgesic prescribing in older adults and associations with clinical and service outcomes [version 1; peer review: awaiting peer review]. NIHR Open Res 2026, 6:160 (https://doi.org/10.3310/nihropenres.14326.1)
NOTE: If applicable, it is important to ensure the information in square brackets after the title is included in all citations of this article.
Export Citation
Sciwheel
EndNote
Ref. Manager
Bibtex
ProCite
Sente
▬
✚
Study Protocol
Helping to Optimise Pain control in the Elderly experiencing Analgesic Overprescribing (HOPE-AO): protocol for a descriptive epidemiological analysis of patterns of analgesic prescribing in older adults and associations with clinical and service outcomes
[version 1; peer review: awaiting peer review]
PUBLISHED 15 Oct 2026
OPEN PEER REVIEW
REVIEWER STATUS
This article is included in the Programme Development Grants gateway.
Abstract
Plain Language Summary
Corresponding author:
Stuart G. Spicer
Competing interests:
No competing interests were disclosed.
Grant information:
This project is funded by the National Institute for Health and Care Research (NIHR) (Grant Reference Number 208337)]. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Copyright:
© 2026 G. Spicer S et al.
This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
How to cite: G. Spicer S, A Payne R, Schofield P et al. Helping to Optimise Pain control in the Elderly experiencing Analgesic Overprescribing (HOPE-AO): protocol for a descriptive epidemiological analysis of patterns of analgesic prescribing in older adults and associations with clinical and service outcomes [version 1; peer review: awaiting peer review]. NIHR Open Res 2026, 6:160 (https://doi.org/10.3310/nihropenres.14326.1)
First published: 15 Oct 2026, 6:160 (https://doi.org/10.3310/nihropenres.14326.1)
Latest published: 15 Oct 2026, 6:160 (https://doi.org/10.3310/nihropenres.14326.1)
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Background
Chronic pain occurs in over 40% of UK community dwelling older adults and is poorly managed.1 About 95 million people across Europe are estimated to have chronic pain,2–5 compounded by inequalities in access to treatment. A large-scale survey across 15 European countries3 high shows high levels of pain medication prescribing: NSAIDs (44%); weak opioids (23%); paracetamol (18%); and strong opioids (5%). Also, 40% have inadequate management of their pain. Whilst this survey focussed across all age groups, a recent overview by the International Association for the Study of Pain highlights disparities in effective pain management which include age, along with sex/gender, ethnicity and socioeconomic status.6
Pain prescribing in older adults presents many challenges, including a reduction in the effectiveness of medication in the long term and increased tolerance, along with multimorbidity and polypharmacy (prescribing of multiple drugs).7,8 Overprescribing of pain medication is a substantial concern, and a key DHSC report9 has highlighted the problem of overprescribing of dependence forming pain medications, including pain medication. Overprescribing of pain medication, including opioids, is particularly prevalent in older adults.10,11 Importantly, there is limited understanding about the older adult population and the best approach to pain management.
In addition to overprescribing, analgesic medication it is often co-prescribed with other medications, including those used to treat psychological distress (e.g. antidepressants, z-drugs, and benzodiazepines). Older adults are prescribed a median of seven medications and polypharmacy is strongly associated with increased mortality and morbidity.12 Notably, several medications are prescribed for both pain and distress (e.g. tricyclic antidepressants and gabapentinoids), and pain and distress syndromes together represent a range of related disturbances of brain-body functioning13 that share underlying social causes including trauma and adversity.14 Such medications share distinct but related mechanisms of tolerance and withdrawal.15 The European Pain Federation has issued a position paper on the use of these drugs in the general population, but no such paper has been produced for older adults.11 There is therefore a strong rationale to consider wider co-prescribed psychotropic medication, alongside the prescribing of chronic pain medication.
A small survey of 100 adults 65+ years regarding their medication intake revealed that 85% would be willing to stop if recommended by their GP, 62% would be willing to take fewer medications, and 41% would want follow-up by a specialist polypharmacy clinic.12 A scoping review16 highlighted how both patients and health care professionals were anxious about stopping medications, even though patients felt overwhelmed by the amount they were being prescribed.
A recent report by Chris Whitty11 suggests that overprescribing in older adults needs to be reduced, to enable improved quality of life. There is limited understanding of chronic pain prescribing practices and patterns in older adults, for medications including opioids, non-steroidal pain medication (NSAIDs), low dose tricyclics, paracetamol and gabapentinoids. The most appropriate framework for person-centred medication deprescribing, optimisation and pain management in older adults is yet to be established. To decide the focus of such a programme we first need to know more about the population itself and current service use. This is the aim of the package of work detailed in this protocol. This work package forms part of a wider mixed-methods project, investigating chronic pain prescribing in older adults.17,18
Objectives, aims and rationale
The overarching aim of this exploratory work package is to explore chronic pain prescribing in older adults, to develop a good understanding of analgesic prescribing patterns in the context of various clinical, therapeutic and sociodemographic factors. This will help us to understand which patients are most likely to benefit from targeted interventions as well as shaping potential intervention strategies.
Objective 1: Understand which sociodemographic (e.g. age, gender, deprivation, rurality, ethnicity) and clinical (e.g. mental health, multimorbidity, frailty, co-prescribing) factors are associated with prescribing of key analgesic drug classes (i.e. weak opioids, strong opioids, NSAIDs, paracetamol, gabapentinoids, low-dose tricyclics), combinations of these drug classes, and changes over time. This includes pattens of co-prescribing of medication for common mental health conditions and psychological distress.
Objective 2: Investigate the association between the aforementioned analgesic drug classes (including clusters of drug classes and temporal changes), and important clinical and health service outcomes (particularly adverse outcomes).
Methods
Patient and public involvement and engagement
Patient and public involvement and engagement (PPIE) has been central to the design and development of this proposal, and we have held several PPIE workshops. We have put together a cross-site diverse PPIE panel of ten people, overseen by a dedicated lived experience PPIE lead. The PPIE panel will meet monthly; responsibilities include solving problems, reviewing progress, sharing ideas and expertise, planning activities, and reviewing and interpreting results. We have already consulted with the DeStress study (University of Exeter),19 who will have input into PPIE activities. We will formally evaluate PPIE contributions using the NIHR PiiAF. Our PPIE Lead has lived experience of chronic pain prescribing, is the Founder & Chair of Trustees for Burning Nights CRPS Support, and is Chair of the Expert Patient and Carer Committee at the British Pain Society.
Design and study plan
This 18-month study utilises separate, parallel quantitative and qualitative work packages (WPs), integrating the findings at the end to inform the design of a future NIHR funded programme grant application. The first quantitative WP (the focus of this protocol) comprises an analysis of the Clinical Practice Research Datalink (CPRD) database,20 to get a clearer understanding of patterns and relationships of chronic pain prescribing within older (aged 65+) adults in England. The second qualitative WP comprises interviews to explore the perceptions and experiences of older adults prescribed pain medication, along with informal carers, and healthcare professionals, including prescribers. The findings of these two work packages will stand alone as important contributions to the literature, but in the final months of the project, we will also integrate the findings of both into a plan for a future programme of research, to codevelop a framework for optimal prescribing and pain management. The qualitative work package is detailed in a separate linked protocol paper, along with details about cross-cutting themes and broader project aims and outputs.18
The analyses of the CPRD database employs a retrospective cohort study design. This allows the nature of analgesic prescribing to be examined in relation to current sociodemographic and clinical factors (Objective 1), and also in relation to subsequent service utilisation post-index date (Objective 2).
Study population
Older adults aged 65+ years on the study index date, registered with at least 12 months of prior data, and who have received at least one prescription for analgesia (defined under outcomes section) in the previous 12 months. We will define our population based on prescribing in 2023, to ensure a contemporary population whilst allowing adequate follow-up time. There is no comparison group.
Outcomes
Outcomes to be measured are split into two categories:
• Prescribing outcomes: rate of analgesic prescribing (by chronicity, and by drug grouping/class/dose).
• Clinical and service outcomes: General Practice doctor, nurse, pharmacist and physiotherapist consultation rates; cascade co-prescribing (e.g. laxatives to mitigate opioid-related constipation); hospital admission rates (Hospital Episode Statistics; HES), including for specific analgesic-related adverse effects (e.g. GI bleed/ulceration with NSAID); referrals to key services (e.g. physiotherapy, pain clinics).
Feasibility counts
We have not conducted a formal feasibility count within CPRD for the current study. In a previous study of disease prevalence in CPRD, we found 10% of adults were taking long-term analgesic medication.21 Analgesic prescribing is also far commoner in older adults (approximately 5-fold greater in over-65 s compared with adults under 45 years. Approximately 20% of the population is analgesic prescribed in people aged over 65 years.22 If only currently registered patients in the CPRD are considered, this provides a population of over 600,000 older patients prescribed analgesics. National primary care dispensing data [openprescribing.net] suggests similar rates of prescribing of non-opioid analgesics and opioid analgesics.
Sample size considerations
Our sample size decisions are based on the precision of estimates of prevalence, rather than for the purposes of pre-specified hypothesis testing. A sample of over 600,000 older adults will allow precision of the overall rate of prescribing of rarely (0.1%) and commonly (10%) used drugs to be estimated with 95% confidence intervals of 0.09–0.11% and 9.9–10.1% respectively. In addition, this still allows precise estimates of such prescribing rates in subgroups only one-tenth of the total population (0.07–0.13% and 9.7–10.3% respectively). Relevant service outcomes can be estimated with similar precision.
Planned use of linked data
Linked HES Admitted Patient Care (APC) data for unplanned hospital admissions will be used to identify important, relevant adverse clinical outcomes (e.g. acute kidney injury, bleeding, delirium, fractures and falls) to address Objective 2. Adverse drug reactions will also be captured through relevant codes within both primary care and HES data. This is to understand risks and potential harms associated with prescribing of pain medication in older adults, and to inform our future recommendations for safer chronic pain management (beyond this CPRD work package), including potential tapering and deprescribing, if appropriate.
Linkage to ethnicity, Index of Multiple Deprivation (IMD) and rural-urban status are required, as socioeconomic factors such as socio-economic deprivation and access to care are likely to influence patient health-seeking behaviour and prescriber decision-making, resulting in potential variation in prescribing patterns, and so will need careful consideration in any future intervention design. The ethnicity linkage through HES will improve the quality of ethnicity status for analysis, beyond use of primary care data alone.
Exposures and covariates
Sociodemographic factors will be determined at the index date, and will include age category (six 5-year groups, up to 90+); sex/gender; IMD (patient-level quintiles); ethnicity; and rurality (including coastal status if possible).
Clinical factors will be determined at the index date, and will include Cambridge Multimorbidity Score (CMS; 20-item general outcome version); electronic Frailty Index (eFI), relevant mental health (based on CMS components, e.g. anxiety, depression, severe mental illness, substance misuse); specific long-term pain conditions (based on CMS components or Health Data Research UK [HDRUK] phenotypes, e.g. musculoskeletal conditions [defined by HDRUK phenotype PH847], endometriosis [HDRUK PH162], migraine [CMS], neuropathic pain/neuralgia [HDRUK PH235, PH549, PH325]); functional disorders (e.g. irritable bowel syndrome [CMS], fibromyalgia [HDRUK PH138, PH866]); and other risk factors (e.g. previous falls [clinicalcodes.org res84/res202], respiratory disorders [based on CMS components], chronic kidney disease [CKD] [based on eGFR, stage 3–5], exploration of past trauma [if adequately coded]). Other prescribing covariate factors to be considered will include polypharmacy (drug count in past 3 months, based on level-4 BNF codes) and relevant specific co-prescribing (e.g. antidepressants [BNF 4.3], benzodiazepines [BNF 4.1.1.0.L/N/P/Q/R/T, 4.1.2.0.A/G/K/P/T], z-drugs [BNF 4.1.1.0.W/Y/Z]) and cascade prescribing (e.g. laxatives [CMS codes], proton pump inhibitors [BNF 1.3.5]). Established codesets from our existing resources, HDRUK or similar, will be used to minimise data preparation time. The aforementioned code definitions may be adapted following clinical review. Clinical factors will be defined using primary care data only (hospital morbidity data will not be used, due to inconsistencies in recording of secondary diagnoses in HES, and to ensure consistency with CMS which is defined based on primary care data only).
Prescribing outcomes
Prescribing will be evaluated primarily based on grouping by BNF categories, including: 4.7.1 (paracetamol, sub-categorised by compound-opioid status), 4.7.2 (opioids, sub-categorised by strength), 4.8.1 (gabapentin, pregabalin), 4.3.1 (low dose amitriptyline and nortriptyline), and 10.1.1 (NSAIDs). We anticipate the following key outcomes for each drug group or cluster (other time frames will be explored):
• “Current”: > = 1 prescription in the last 3 months (also comparing > = 1 prescription in the last 6 weeks for sensitivity)
• “Chronic AND Current”: > = 6 prescriptions in the last 12 months, with > = 1 of these issued in the last 3 months (> = 1 prescription in the last 6 weeks for sensitivity; also comparing 4 and 8 prescriptions in the last 12 months for sensitivity)
The time frames and prescribing during these periods are based on clinical expert opinion. However, given uncertainty over the ideal cut-points, alternatives will be examined as noted above (e.g. 4+ or 8+ issues within 12 months; current prescriptions defined within 6 weeks). Given the nature of the proposed analysis, we envisage undertaking exploratory work to determine whether there is value in additionally grouping analgesic medications by strength. Due to the well-recognised challenges of reliably determining dose using CPRD, we anticipate using prescription strength primarily based on drug preparation alone, although exploration of prescription strength based additionally on daily dose will also be carried out.
Clinical and service outcomes
Service use will be assessed using GP consultation rates (face-to-face or phone), unplanned hospital admission rates (HES APC) and referrals to key services (e.g. physiotherapy, pain clinics). Linked HES data for unplanned hospital admissions will be used to identify important, relevant adverse clinical outcomes (e.g. acute kidney injury [HDRUK PH37], bleeding [PH1026, PH1648], delirium [PH147], falls and fractures [clinicalcodes.org res84/res202]). Adverse drug reactions will also be captured through relevant codes within both primary care and HES data.
Analyses
This retrospective study is descriptive, exploratory, and hypothesis generating, and will focus on describing patterns of analgesia prescribing in the study population, and the association of prescribing with relevant service outcomes.
Key elements of data management:
• Study period: Due to the potential for seasonal variation in prescribing behaviour and service utilisation, we propose using a random index date occurring during the calendar year 2023 for each patient. This will allow for a 1-year follow-up (i.e. to end 2024). If necessary, we will use earlier dates to ensure complete HES data are available for the follow-up period.
• Prescription processing: Prescribing data will be processed by grouping each product code by drug substance. Exploratory work will be undertaken to additionally examine grouping drug codes by strength or dose. Counts for a defined group will be based on number of corresponding issue dates occurring in a specified time-period. Drug classification will be determined by clinical experts within the wider research team.
• Co-morbidities: Co-morbidities will be determined relative to the index date, based on definitions used by the CMS or (for non-CMS morbidities) as any relevant diagnostic SNOMED code occurring prior to the index date.
Four principal analyses will be conducted:
1. Rates of prescribing: Simple descriptive statistics will be used to describe rates of chronic and current (as defined earlier) prescribing of key analgesic drug groups. Latent class analysis (LCA) will be used to assign each patient to an analgesic prescribing cluster, based on the probabilistic approach we have used previously23; two LCA analyses will be conducted, one for chronic and one for current prescribing. Denominator data will be used to determine population rates.
2. Trends in prescribing: To provide insights into change in prescribing over time, simple descriptive statistics will be used to describe rates of current (last 3 months) prescribing, based on the nature of prior prescribing of the same drug (i.e. no prior prescription; same substance with same, lower or higher strength; different substance but same class). Prior prescribing will be determined one year previously (i.e. 12–15 months pre-index date); alternative time frames will be explored. Latent class growth analyses (LCGA) will be used to identify clusters of prescribing change over time. Longer-term prescribing trends will be explored using time series analysis (e.g. ARIMA and STL, GAM’s, and mixed-effects models, e.g. to detect increases/decreases in prescribing and co-prescribing, along with trends related to sociodemographic and clinical covariates).
3. Association of prescribing patterns with sociodemographics and clinical factors: Multi-level Poisson regression (or negative binomial model, with/without zero-inflation, as appropriate) will be used to model the association between rates of prescribing of different analgesic groups and prescribing clusters currently, long-term and over time, with sociodemographic and clinical covariates of interest. Models will include random intercepts to account for clustering of prescribing behaviours by practice.
4. Association of prescribing patterns with subsequent service use: Multi-level regression (Poisson, logistic or Cox, as appropriate) will be used to model the association between unplanned hospital admission, occurrence of adverse drug events, rate of GP consultation, and relevant referrals, with analgesic prescribing/clusters of prescribing.
Risk mitigation plan for small cell counts: With respect to publication, the characteristics of small groups (i.e. <5 patients) will only be reported using one or two broad categories (e.g. a 5 or 10 year age range, as opposed to single years), and will not combine multiple characteristics (e.g. age + sex/gender + ethnicity + area) that could risk disclosure. During conduct of the study, we will not be matching to any other patient identifiable data that would allow re-identification of patients.
Addressing confounding effects
To control for any confounders relating to the association between prescribing patterns and service use, we will adjust for key patient and clinical factors potentially associated with both service use and prescribing (e.g. diagnoses, age, sex/gender, ethnicity; along with clustering by practice). We acknowledge that this still does not allow us to infer a causal association. However, we do not intend to imply causation, and the analysis will still provide critical insights into the nature of the association between prescribing and service use which will be invaluable to help us design service-level interventions.
Addressing missing data
We will use patient records where only complete data are available for age, sex and other relevant sociodemographics, as we anticipate completeness of these data to be extremely high (>99%) and any bias from including only complete data will thus be minimal. Where ethnicity is missing, this will be coded as such in the analyses. No missing data analysis is planned for prescribing or clinical data, as the absence of a relevant code is assumed to correctly indicate that the corresponding drug or condition is not present. We do not intend to include clinical factors such as smoking or BMI for which missingness might be relevant. For missing prescribing dosing or quantity data, we will impute values based on the approach we have used previously.24
Limitations
Prescribing is likely to be highly complete, given the predominance of electronic prescribing in UK general practice. However, we recognise that dosing information, in particular, may be missing or subject to inaccuracies. This is acknowledged above.
Ethnicity is likely to have higher rates of missingness as this is an older patient population.
Our analyses will not capture over-the-counter analgesic use, although our interest is in prescribed drugs irrespective of concomitant non-prescription use.
Definitions of morbidities are dependent on accurate coding by general practitioners, and it possible that rates of recording may be impacted by factors which also affect propensity to prescribe analgesics.
We are unable to capture non-GP consultations, such as private physiotherapists or community pharmacy.
No causal inference will be made.
Dissemination
We anticipate at least one publication of findings and a conference presentation. We will hold two workshops at the end of the project to disseminate findings and plan subsequent research. We will use existing networks, social media and a website to disseminate findings, plus mainstream media where feasible to maximise interest. We have good links with subject-specific networks to reach a pharmacy audience and will engage key stakeholders throughout the project.
This study may impact policies. For example, the Scottish Intercollegiate Guidelines Network (SIGN) is keen to address the issue of deprescribing in their guidance. The National Guidelines for the management of pain, recommend caution in prescribing any analgesics for long periods in older adults; this study may influence implementation of these guidelines.
This project will be of interest to specialist groups, including (but not limited to) the British Pain society, British Geriatrics Society, Cornwall & Devon Pain Café network, and the Health Innovation Network.
In terms of publication, we will mitigate any potential risk of patient identification due to small numbers in some potential sub-groups. Such small groups will not be reported, and during conduct of the study we will not be matching to any patient identifiable data that would allow deanonymisation.
Ethical Approval and informed consent
Ethical approval for the wider study, including the qualitative components, was obtained from the Health Research Authority and Solihull Research Ethics Committee (IRAS ID: 344044). Approval for the Clinical Practice Research Datalink (CPRD) work package analyses, described in this protocol, was obtained through the CPRD Research Data Governance process via the eRAP system,25 in accordance with CPRD governance requirements (eRAP reference number 24_004610). All research activities are conducted in accordance with the principles of the Declaration of Helsinki. Individual patient consent is not required for the work package described in this protocol, as it uses routinely collected pseudonymised/anonymised data obtained through the CPRD under its established governance framework. However, informed consent for participation in the wider study work packages is obtained for anyone participating in these research activities (i.e. qualitative interviews and health economics questionnaires).
Funding and acknowledgements
This study has been funded by the NIHR programme development grant 208337. Stuart G. Spicer, Pat Schofield and Richard Byng were additionally supported by the National Institute for Health and Care Research Applied Research Collaboration South West (NIHR ARC SW). The views expressed are those of the authors and not necessarily those of the NIHR.
Download
Export To
CITE
how to cite this article
G. Spicer S, A Payne R, Schofield P et al. Helping to Optimise Pain control in the Elderly experiencing Analgesic Overprescribing (HOPE-AO): protocol for a descriptive epidemiological analysis of patterns of analgesic prescribing in older adults and associations with clinical and service outcomes [version 1; peer review: awaiting peer review]. NIHR Open Res 2026, 6:160 (https://doi.org/10.3310/nihropenres.14326.1)
NOTE: If applicable, it is important to ensure the information in square brackets after the title is included in all citations of this article.
track
receive updates on this article
Track an article to receive email alerts on any updates to this article.
Open Peer Review
Current Reviewer Status:
AWAITING PEER REVIEW
AWAITING PEER REVIEW
Key to Reviewer Statuses
VIEW
HIDE
ApprovedThe paper is scientifically sound in its current form and only minor, if any, improvements are suggested
Approved with reservations
A number of small changes, sometimes more significant revisions are required to address specific details and improve the papers academic merit.
Not approvedFundamental flaws in the paper seriously undermine the findings and conclusions
Alongside their report, reviewers assign a status to the article:
Approved - the paper is scientifically sound in its current form and only minor, if any, improvements are suggested
Approved with reservations -
A number of small changes, sometimes more significant revisions are required to address specific details and improve the papers academic merit.
Not approved - fundamental flaws in the paper seriously undermine the findings and conclusions
Competing Interests Policy
Provide sufficient details of any financial or non-financial competing interests to enable users to assess whether your comments might lead a reasonable person to question your impartiality. Consider the following examples, but note that this is not an exhaustive list:
Examples of 'Non-Financial Competing Interests'
- Within the past 4 years, you have held joint grants, published or collaborated with any of the authors of the selected paper.
- You have a close personal relationship (e.g. parent, spouse, sibling, or domestic partner) with any of the authors.
- You are a close professional associate of any of the authors (e.g. scientific mentor, recent student).
- You work at the same institute as any of the authors.
- You hope/expect to benefit (e.g. favour or employment) as a result of your submission.
- You are an Editor for the journal in which the article is published.
Examples of 'Financial Competing Interests'
- You expect to receive, or in the past 4 years have received, any of the following from any commercial organisation that may gain financially from your submission: a salary, fees, funding, reimbursements.
- You expect to receive, or in the past 4 years have received, shared grant support or other funding with any of the authors.
- You hold, or are currently applying for, any patents or significant stocks/shares relating to the subject matter of the paper you are commenting on.
Stay Updated
Sign up for content alerts and receive a weekly or monthly email with all newly published articles
Register with NIHR Open Research
Already registered? Sign in
Comments on this article Comments (0)