Dr Maria Boyce, UCC 2009
Dr Maria Boyce
The effectiveness of using patient-reported outcome measures as quality improvement tools
Institution: UCC, 2009 cohort
Supervisors: Prof. John Browne (UCC)
To investigate the value of using PROMs as quality improvement tools.
Two systematic reviews were undertaken. The first reviewed the quantitative literature on the impact of PROMs feedback and the second reviewed the qualitative literature on the use of PROMs in practice. These reviews informed the focus of the primary research. A cluster randomised controlled trial (PROFILE) examined the impact of providing peer benchmarked PROMs feedback to consultant orthopaedic surgeons on improving outcomes for hip replacement surgery. Qualitative interviews with surgeons in the intervention arm of the trial examined the view of and reactions to the feedback.
The quantitative review of 17 studies found weak evidence to suggest that providing PROMs feedback to professionals improves patient outcomes. The qualitative review of 16 studies identified the barriers and facilitators to the use of PROMs based on four themes: practical considerations, attitudes towards the data, methodological concerns and the impact of feedback on care. The PROFILE trial included 11 surgeons and 215 patients in the intervention arm, and 10 surgeons and 217 patients in the control arm. The trial found no significant difference in the Oxford Hip Score between the arms (-0.7, 95% CI -1.9-0.5, p=0.2). Interviews with surgeons revealed mixed opinions about the value of the PROMs feedback and the information did not promote explicit changes to their practice.
It is important to use PROMs which have been validated for the specific purpose of performance measurement, consult with professionals when developing a PROMs feedback intervention, communicate with professionals about the objectives of the data collection, educate professionals on the properties and interpretation of the data, and support professionals in using the information to improve care. It is also imperative that the burden of data collection and dissemination of the information is minimised.
Dr Barbara Clyne, RCSI 2009
Dr Barbara Clyne
Effectiveness of medicines review with web-based pharmaceutical treatment algorithms in reducing potentially inappropriate prescribing in older people in primary care: a cluster randomized trial (OPTI-SCRIPT study)
Institution: RCSI, 2009 cohort
Supervisors: Prof. Tom Fahey (RCSI), Prof. Susan Smith (RCSI), Dr. Marie Bradley (RCSI)
Potentially inappropriate prescribing (PIP) is common and can result in increased morbidity, adverse drug events and hospitalisations. In 2007, 36% of Irish people aged 70 years or over received at least one potentially inappropriate prescription, with an associated expenditure of €45 million. This thesis documents the design, conduct, and evaluation of a cluster randomised controlled trial (RCT) in Irish primary care focused on reducing PIP in older patients (> 70 years).
The intervention development was guided by the Medical Research Council (MRC) framework for the development and evaluation of complex interventions. Literature and information from experts were combined to define the main components of the intervention which was then tested with five GPs during the pilot and evaluated using qualitative interviews
During the OPTI-SCRIPT trial 21 practices and 196 patients were allocated to intervention (11 practices, 99 patients) or control (10 practices, 97 patients) using minimisation. Intervention participants received a multifaceted intervention incorporating academic detailing with a pharmacist, medicines review with web-based pharmaceutical treatment algorithms that provided recommended alternative treatment options, and tailored patient information leaflets. Control practices delivered usual care and received simple patient-level PIP feedback. Primary outcomes were the proportion of participants with PIP and the mean number of potentially inappropriate prescriptions per group. Secondary outcomes included drug-specific outcomes, well-being and beliefs about medication. Routinely collected national prescribing data were analysed as a contemporary national comparison.
During the mixed methods process evaluation, semi-structured interviews with participating GPs and a purposive sample of patients in the intervention group were conducted. Interviews were transcribed verbatim and the data analysed using a thematic analysis. Data were also collected from practice questionnaires, website activity and researcher logs.
All 21 practices and 190 patients were followed up. At follow-up there was a statistically significant difference between groups. For patients in the intervention group, the odds of not having PIP were 3 times higher than the odds in the control group (OR 3.1 95% Cl 1.4, 6.5). The average PIP in the intervention group was also significantly lower than control (0.70 v 1.19, p=0.004). In the control group, the incidence rate for PIP was 1.42 times the incidence rate for the intervention group. The intervention had a statistically significant effect in reducing proton pump inhibitor prescribing but not on the other secondary outcomes. Exploratory analysis of national prescribing data highlighted that PIP was increasing during the study period and that the control group did not act differently to national trends.
The process evaluation highlighted that requiring GPs to be involved in patient recruitment was time consuming. GPs received a standardised academic detailing session but delivered the intervention differently. Patient information leaflets were not used at all. Participants identified important barriers and facilitators to the future effectiveness of such an intervention, with time being a barrier reported by both GPs and patients. Overall, the intervention was positively received with participants referring to the value of medication reviews to improve prescribing practices and reduce unnecessary medications in older patients.
The study found that a multifaceted intervention incorporating academic detailing with a pharmacist, medicines review with web-based pharmaceutical treatment algorithms that provide recommended alternative treatment options, but not tailored patient information leaflets, was effective in reducing PIP, particularly for proton pump inhibitors, but had no effect on other secondary outcomes. The process evaluation highlighted that the intervention was acceptable to the majority of participants.
Dr Margaret Curtin, UCC 2009
Dr Margaret Curtin
Can population-level variation in Early Childhood Development be measured in a manner that supports service planning and intervention?
Institution: UCC, 2009 cohort
Supervisors: Prof. Ivan Perry (UCC), Prof. Anthony Staines (DCU)
The Early Development Instrument (EDI) is a population-level measure of five developmental domains at school-entry age. The overall aim of this thesis was to explore the potential of the EDI as an indicator of early development in Ireland.
A cross-sectional study was conducted in 47 primary schools in 2011 using the EDI and a linked parental questionnaire. EDI (teacher completed) scores were calculated for 1,344 children in their first year of full-time education. Those scoring in the lowest 10% of the sample population in one or more domains were deemed to be ‘developmentally vulnerable’. Scores were correlated with contextual data from the parental questionnaire and with indicators of area and school-level deprivation. Rasch analysis was used to determine the validity of the EDI.
Over one quarter (27.5%) of all children in the study were developmentally vulnerable. Individual characteristics associated with increased risk of vulnerability were being male; under 5 years old; and having English as a second language. Adjusted for these demographics, low birth weight, poor parent/child interaction and mother’s lower level of education showed the most significant odds ratios for developmental vulnerability. Vulnerability did not follow the area-level deprivation gradient as measured by a composite index of material deprivation. Children considered by the teacher to be in need of assessment also had lower scores, which were not significantly different from those of children with a clinical diagnosis of special needs. all domains showed at least reasonable fit to the Rasch model supporting the validity of the instrument. However, there was a need for further refinement of the instrument in the Irish context.
This thesis provides a unique snapshot of early development in Ireland. The EDI and linked parental questionnaires are promising indicators of the extent, distribution and determinants of developmental vulnerability.
Dr Bairbre Meaney, TCD 2009
Dr Bairbre Meaney
From the inside out: an idiographic account of the policy making process that produced the Irish work place smoking ban
Institution: TCD, 2009 cohort
Supervisors: Prof. Charles Normand (TCD), Dr. Elaine Byrne (RCSI), Prof. Imelda Coyne (TCD)
Abstract: Not available
Dr Lisa Mellon, RCSI 2009
Dr Lisa Mellon
Time to hospital presentation in acute ischaemic stroke
Institution: RCSI, 2009 cohort
Supervisors: Prof. Anne Hickey (RCSI), Prof. Frank Doyle (RCSI), Prof. David Williams (RCSI)
For an ischaemic stroke, revascularisation treatment with thrombolysis is associated with reductions in morbidity and mortality. Despite its proven benefits, thrombolysis therapy is under-utilised in clinical practice with patient delay in presenting to hospital with symptoms identified as the leading barrier to thrombolysis administration. Population-focussed behaviour change interventions such as mass media interventions have demonstrated impacts on population stroke knowledge and intentions to seek medical help when symptoms occur. However their impact on behaviour during stroke onset is not established.
The programme of work in this dissertation sought to: determine the success of interventions to reduce patient delay with stroke symptoms (systematic review); investigate the behavioural impact of an Irish mass media intervention for stroke awareness (interrupted time-series analysis); examine the complex factors involved in decision-making during stroke onset (cross-sectional study); and examine the long-term impact of hospital arrival time on patient outcome (longitudinal study).
The results highlighted the complexity of changing help-seeking behaviour during stroke onset. The FAST campaign is associated with demonstrated changes in help-seeking behaviour for stroke, however there is scope for further improvement. A number of implications for stroke education strategies and future research are documented. The potential clinical benefits of media interventions such as FAST are far-reaching and the challenge remains in translating knowledge improvements and correct behavioural intentions to appropriate behaviour, where people act on the message. Minimising pre-hospital delay in this way serves to reduce the potentially devastating impact of stroke.
Dr Patrick Moore, TCD 2009
Dr Patrick Moore
Counting the Time Lived or the Time Left? Age, Proximity to Death and Prescription Expenditures
Institution: TCD, 2009 cohort
Supervisors: Prof. Charles Normand (TCD), Prof. Kathleen Bennett (RCSI)
This thesis set out to investigate whether ageing is a surrogate measure for proximity to death in the older population when estimating medication expenditures. The medication expenditures used were the ingredient cost of medications for the public health system. Concerns about the long term sustainability of health care expenditures and in particular prescribing expenditures has become an important policy issue in most developed countries. Previous studies suggest that proximity to death (PTD) has a significant effect on total health care expenditures, with its exclusion leading to an overestimation of likely growth. There are limited studies of pharmaceutical expenditures taking PTD into account.
This thesis uses individual level dispensing data for the New Zealand and Irish population aged 70 years or more. The period used, 2006 – 2009, provides a unique experiment in Ireland when individuals aged 70 years or more were entitled to free health care and medications paid for by the state regardless of income, illness or socioeconomic status. Given this eligibility it was possible to use dispensing data automatically collected at pharmacies for reimbursement purposes as a national cohort.
A case control methodology is used to examine individual expenditure and medication use for a 12-month period for decedents (cases) and survivors (controls). The hypothesis was tested using a random effects two-part model, consisting o f a Probit regression to identify the probability of expenditure in any given month and a generalised linear model (GLM) to regress the positive expenditures. New Zealand was used as a comparator country given the similarities in demographics, socioeconomic circumstances and access to free medication.
The impact of proximity to death on prescription expenditure is not as dramatic as studies reporting on Acute and/or long term care. Decedents typically expenditure up to twice as much as their surviving counterparts in terms of medication expenditure in both countries. In Ireland the average monthly medication expenditure for an individual in the last year of life is €2,107, compared to €1,180 for a similar survivor. Decedents consistently spend and use more medications up to three years before death. The data show a neutral effect for age once PTD is included which has a positive and statistically significant impact on prescribing expenditures. Medium term expenditure projections are overestimated when PTD is not taken into account.
A similar effect was reported for New Zealand the 12-month decedent to survivor mean expenditure ratio was 1.95, 2.09 for males and 1.82 for females. The additional expenditure of dying in terms of prescription drugs decreases with age, with those who die at 90 years of age or older consuming fewer drugs on average and having a lower mean expenditure than those who died in their 70s and 80s. The following variables were found to have a decreasing effect on the mean monthly prescription expenditures, a reduction of 2.2% for each additional year of age, 4.2% being in the Maori ethnic group and 7.8% for Pacific Islanders. Increases in monthly expenditure were associated with being a decedent 32.1%-62.6% (depending on month), being of Asian origin 16.2% or a male 12.6%.
Given the variance reported between survivors and decedents, to improve accuracy future projections should include PTD in their models. Policies targeted at reducing expenditures should not focus on age but on ensuring appropriate and cost effective prescribing especially towards the end of life.
Dr Caitriona O’Neill, TCD 2009
Dr Caitriona O’Neill
Costs and outcomes of cancer care in older adults: Population-based analysis
Institution: TCD, 2009 cohort
Supervisors: Prof. Martin Henman (TCD)
Abstract: Not available
Dr Miriam O’Shea, TCD 2009
Dr Miriam O’Shea
Comorbidity and Type 2 Diabetes: prevalence, type and impact on cost and health-related behaviours
Institution: TCD, 2009 cohort
Supervisors: Prof. Kathleen Bennett (RCSI), Dr. Mary Teeling (TCD), Prof. Michael Barry (TCD)
Comorbidity is the co-existence of one or more additional conditions in patients with a specified medical condition. It is highly prevalent in patients with diabetes and may be related (concordant) or unrelated (discordant) to diabetes. Few studies in Ireland have investigated the presence or impact comorbidity in patients with Type 2 Diabetes Mellitus (T2DM). International studies have shown that the presence of comorbidity in patients with diabetes has negative implications in terms of the patient, economy, and society. Until recently however, the majority of these studies focused, only, on the presence and/or impact of concordant comorbidity.
The overall aim of this thesis was to investigate the prevalence, type, and cost of both concordant and discordant comorbidity and its impact on adherence and self-care in patients with T2DM in Ireland by: (i) estimating the prevalence, type and ingredient drug cost of comorbidity associated with medication treated T2DM in both the adult and elderly population, that were eligible for inclusion in the General Medical Services (GMS) scheme in 2010, (ii) examining the influence of comorbidity and region of residence on adherence (the degree to which a patient takes their medication as prescribed) and persistence (the length of time that a patient remains on their medication) to oral anti-hyperglycaemic (OAH) agents in the GMS eligible population and (iii) investigating the impact of comorbidity on diabetes self-care and health-related quality of life (HRQoL) in a diabetes population attending an outpatient department in an Irish hospital. This thesis is composed of five studies.
The first four studies are based on analysis of prescription claims data obtained from the Health Service Executive – Primary Care Reimbursement Service. The first two studies found that, during 2010, the prevalence of treated T2DM in the GMS eligible adult (25 – 64 years) (N=674,026) and elderly (>65 years) (N=445,180) population was 3.1% (N=21,877) and 9.7% (N=43,165) respectively. The median number of comorbid conditions was significantly higher in patients with treated T2DM, compared to those without T2DM, in both the adult (3 vs. 0 p<0.0001) and elderly (5 vs. 3, p<0.0001) patient populations. The higher level of comorbidity in patients with treated T2DM was associated with a significant increase in drug costs compared to those without T2DM with an average increase in costs of €717.29 in the adult and €439.26 in the elderly respectively. Concordant and discordant comorbid conditions were common in adult and elderly patients with T2DM.
The third study examined the influence of comorbidity on medication taking behaviour in new users of any OAH therapy (N=21,280). The overall rate of persistence and adherence to OAH therapy was 74% and 70% respectively at 6 months, and 62.6% and 66.6% respectively at 12 months post initiation. The rate of persistence and adherence to OAH agents was shown to increase with increasing number of comorbid conditions. When comorbidity was stratified by type, patients with only concordant conditions were significantly more likely to be persistent at 6 (OR 1.45, 95% Cl 1.28, 1.65) and 12 months (OR 1.22, 95% Cl 1.09, 1.38) when compared to patients either with no or with both types of comorbidity. In contrast, patients with only discordant conditions were significantly less likely to be persistent at 6 (OR 0.40, 95% Cl 0.35, 0.46) and 12 months post initiation (OR 0.43, 95% Cl 0.38, 0.50). Results for adherence were similar.
The results of the fourth study (N=11,035) suggest there was little variability between the four HSE administrative regions in terms of patient medication taking behaviour. The Southern region had the lowest rates for both non-persistence (HR 0.86, 95% Cl 0.80, 0.94) and non-adherence (OR 0.83 95% Cl 0.74, 0.93). The presence of any type of comorbidity (versus no comorbidity) was shown to be associated with a lower risk of both non-persistence and non-adherence to OAH therapy in all regions. The results presented in fifth study relate to the responses (N=159) received from an anonymised cross-sectional questionnaire that was sent to 498 T2DM patients that had attended an outpatient diabetes clinic in St James’s Hospital Dublin. In relation to diabetes self-care, type of comorbidity was significantly associated with physical activity. Respondents with no comorbid conditions had the highest median value for physical activity, while those with both types of comorbidity had the lowest. In terms of HRQoL, the presence of discordant comorbidity either on its own or in combination with concordant comorbidity was also associated with a significant decline in EQ-5D score compared to no complications / comorbid conditions. The presence of concordant comorbidity on its own was not found to significantly
affect EQ-5D scores.
The studies presented in this thesis are currently among the few in Ireland to examine the presence and / or impact of comorbidity in patients with T2DM. The results of the first two studies provide baseline prevalence for concordant and discordant comorbid conditions in the adult and elderly GMS eligible population with treated T2DM. The results of these studies suggest that both types of comorbidity are highly prevalent in these patients. The drug costs associated with the treatment of both types of comorbidity was found to be considerably higher in patients with T2DM when compared to other patients. The economic cost of treating concordant conditions in patients with diabetes has been recognised in the international literature. The results presented in this thesis add to the existing literature by presenting costs relating to the treatment of discordant comorbidity. Previous international studies have also suggested that the presence of comorbid conditions may have a deleterious impact on patient medication taking behaviour and HRQoL in patients with T2DM. The results presented in this thesis add to this literature by showing that the presence of discordant comorbid conditions is associated with lower persistence and adherence to OAH agents and reduced HRQoL.