Article Text
PDF
Abstract
Introduction Familial hypercholesterolaemia (FH), an autosomal dominant disorder of lipid metabolism, results in accelerated onset of atherosclerosis if left untreated. Lifelong treatment with diet, lifestyle modifications and statins enable a normal lifespan for most patients. Early diagnosis is critical. This protocol trials a primary care-based model of care (MoC) to improve detection and management of FH.
Methods and analysis Pragmatic cluster intervention study with pre-post intervention comparisons in Australian general practices. At study baseline, current FH detection practice is assessed. Medical records over 2 years are electronically scanned using a data extraction tool (TARB-Ex) to identify patients at increased risk. High-risk patients are clinically reviewed to provide definitive, phenotypic diagnosis using Dutch Lipid Clinic Network Criteria. Once an index family member with FH is identified, the primary care team undertake cascade testing of first-degree relatives to identify other patients with FH. Management guidance based on disease complexity is provided to the primary care team. Study follow-up to 12 months with TARB-Ex rerun to identify total number of new FH cases diagnosed over study period (via TARB-Ex, cascade testing and new cases presenting). At study conclusion, patient and clinical staff perceptions of enablers/barriers and suggested improvements to the approach will be examined. Resources at each stage will be traced to determine the economic implications of implementing the MoC and costed from health system perspective. Primary outcomes: increase in number of index cases clinically identified; reduction in low-density lipoprotein cholesterol of treated cases. Secondary outcomes: increase in the number of family cases detected/contacted; cost implications of the MoC.
Ethics and dissemination Study approval by The University of Notre Dame Australia Human Research Ethics Committee Protocol ID: 0 16 067F. Registration: Australian New Zealand Clinical Trials Registry ID: 12616000630415. Information will be disseminated via research seminars, conference presentations, journal articles, media releases and community forums.
Trial registration number Australian New Zealand Clinical Trials Registry ID 12616000630415; Pre-results.
- primary care
- lipid disorders
- coronary heart disease
This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Statistics from Altmetric.com
Strengths and limitations of this study
To the best of our knowledge, this protocol is the first that focuses on early detection and on the delivery of preventative care and management of familial hypercholesterolaemia (FH) in the primary care setting. It is trialled within the Australian context but builds on the consensus statements of the European Atherosclerosis Society and the International FH Foundation that FH care should ideally take place in the primary care setting.
Our pragmatic approach using existing clinical infrastructure enhances feasibility and sustainability but pre-post intervention comparison is acknowledged as potential limitation.
The general practitioner (GP) and practice nurse (PN) team approach to phenotypic diagnosis and the cascade testing of FH relatives is likely to prove challenging initially from the GP/PN and patient perspective—length of follow-up to determine appropriate management of patients with FH will be limited due to constraints of time and funding.
There is potential for impact on the outcomes and interpretation as patients attending GP practices in Australia are not registered to a single practice—they may change residential address and be lost to follow-up or register at more than one practice with potential for duplication (or loss to follow-up if at a non-participating practice); however, this will be easily identified given the numbers of participants followed up.
The potential for variability in compliance with medications (eg, statins) and adherence to dietary and lifestyle advice across practices and between patients is acknowledged.
Introduction
Familial hypercholesterolaemia (FH) is an autosomal dominant disorder of lipid metabolism resulting in excessively high plasma levels of cholesterol from birth. Left untreated, it can result in premature coronary artery disease (CAD) due to accelerated onset of atherosclerosis.1 For men with FH, the risk of developing CAD before age 50 years is 50% while for women it is 30% at age 60 years.2 3 The accelerated development of atherosclerotic cardiovascular disease (CVD) by 20–40 years as a consequence of inheriting FH means that children and young adults have most to gain from early diagnosis and treatment.4 It is estimated that between 1 in 500 and 1 in 200 persons have heterozygous FH yielding a worldwide population estimate of 20 million cases.5 Of the estimated 45 000 with FH in Australia and New Zealand,6 the vast majority remain undiagnosed and, among those diagnosed, most remain undertreated. However, lifelong treatment with diet and lifestyle modifications together with lipid lowering therapy can reduce the risk of CAD close to the non-FH population.7 Early diagnosis is therefore critical.
The 2011 Australian FH Model of Care (MoC) primarily focused on specialist lipid clinics in tertiary centres to augment the development of clinical services for FH.8 The need to develop a corresponding MoC suitable for the primary care setting is increasingly recognised,8 especially in light of the vastly different disease demographics of patients attending primary and tertiary clinics (according to Bettering Evaluation and Care of Health (BEACH)data, 8.4% of primary care encounters are referred to medical specialists)9 and the persisting low rates of FH detection and treatment in the Australian population.8 10 Primary cares services are the first point of contact with the health system and over 85% of the Australian population attend a general practitioner (GP) at least once annually.11 Thus, primary care teams are ideally placed to play a more active role in the detection and management of unsuspected cases of FH in the community, thereby contributing to addressing the gap that currently exists in the Australian health service.8
The key challenge for primary care is to develop a systematic and sustainable approach to detect index cases for FH in the community. Once this is achieved, the next step is to progress to family cascade testing. There is extensive evidence to show such an approach would be clinically2 8 10 and financially12–14 effective. The need to integrate the central role of primary care with specialist lipid services to optimise the detection and management of FH in the community is recognised.15 16 The provision of a community-based MoC for FH involving a more comprehensive and focused education programme for GPs, practice nurses (PNs) and practice staff, improved communication with cardiologists and pathology providers as well as support from lipid specialists to help manage more difficult high-risk patients with FH and to improve cascade testing of relatives of index cases, has been suggested.16
In a regional Western Australian (WA) primary care setting, three different approaches for FH detection were successfully tested.17 These included (1) a community pathology laboratory database (n=52 200); (2) a workplace-based occupational health assessment process (n=1079), and (3) a general practice patient database (n=41 100) to screen for increased CVD risk—a total population of 94 379. A total of 1316 individuals subsequently had a detailed clinical assessment for FH and 86 participants were identified with clinical FH. Those with Dutch Lipid Clinic Network Criteria (DLCNC)2 scores >5 were offered referral to lipid specialist clinic for further review. Of the 59 individuals assessed by the lipid specialist who underwent DNA testing, 11 (18.6%) were identified to carry FH mutations.17
The study used the existing practice-based medical and nursing services in the regional primary care setting. The findings supported the potential development of an integrated screening programme capable of combining the use of pathology services (including interpretative comments on lipid profile) and the involvement of the primary care team in detection and management.17 The engagement of local GPs and PNs in recalling and following up patients generated greater patient involvement in the screening process. The preferred model used a two-stage process where at-risk patients were identified through raised low-density lipoprotein cholesterol (LDL-c) concentrations that were subsequently flagged by the GP-PN team for review.
Bell et al 18 examined whether individuals with FH could be accurately identified in the primary care setting. They concluded that GPs were able to accurately identify individuals at high or low risk for FH using the DLCNC score,2 thereby augmenting opportunistic FH detection in the community. In addition, by increasing education for the primary care team, the diagnostic accuracy of FH detection in primary care could be enhanced.
With this in mind, Troeung et al 19 designed an electronic data extraction tool (TARB-Ex) which enables routine clinical information to be extracted from general practice patient management software databases (Best Practice Software) and identifies patients at potential risk of FH for follow-up based on the DLCNC score. The extraction tool has shown its capacity to identify patients at increased risk of FH in the primary care setting, comparing favourably with a manual assessment of FH risk by the GP,19 thus facilitating an innovative and time-saving approach to improve both detection and subsequent management.
The current protocol is an innovative primary care-based diagnostic approach using phenotypic criteria as per the DLCNC score rather than more expensive genetic testing. Close relatives of index cases will be targeted for cascade testing and appropriate treatment and lifestyle modifications instituted. Most of this work will be undertaken in the less expensive community setting of general practice. Lipid specialist assistance will be available for more difficult and complex cases but it is anticipated that most of the detection and management of FH will occur in the primary care sector. The MoC is thus both pragmatic and feasible with potential for considerable immediate cost savings to the health sector.
Methods and analyses
Objective
To trial a primary care-based MoC for FH.
Hypothesis
Electronic data extraction from patient records facilitates clinical review to improve the detection and management of FH in general practice.
Study design
A pragmatic, cluster intervention study with pre-post intervention comparisons.
Study setting
General practices in Australia. The study commenced in five practices in WA in July 2016.
Eligibility criteria
Patients aged 18+ years, able to provide informed consent.
Primary outcomes
increase in number of index cases clinically identified;
reduction in LDL-c of treated cases.
Secondary outcomes
increase in the number of family cases detected/contacted (including children);
cost implications of the method of care.
Study procedure
Community involvement
In June 2016, the investigation team worked with WA Consumer and Community Health Research Network20 and included members of the FH Support Group21 and FH Australasia Network22 to organise a ‘Community conversation’ forum23 prior to study commencement in WA. The aim of the forum was to reflect on aspects of the study protocol including the preferred approach to cascade testing of relatives.
Information sessions
Information sessions will be held at the start of each arm of the study for all participating practices. These will be convened as general sessions open to all participating GPs and PNs and supplemented by the investigation team undertaking location visits to each participating practice. The sessions will include background information on FH, the proposed method of care protocol, the process of consenting patients to participate, data to be collected during the research study and suggestions for cascade testing processes. The content for these sessions will be adapted from those previously undertaken.24 25
In addition, GPs and PNs will have the opportunity for one-on-one consultation with the GPs on the research team at each stage of the research protocol. This will allow for specific discussion of individual patients.
Overall plan is outlined in figure 1 and timeline is outlined in table 1.
Study timeline
Steps involved in study protocol. CVD, cardiovascular disease; FH, familial hypercholesterolaemia; GP, general practitioner; PN, practice nurse.
Step 1: assessment of current practice
Qualitative (open-ended) and semiquantitative (scaled) responses will be collected from GPs and PNs on current practice (table 2). This aims to establish a baseline of GP, PN and practice staff awareness and GP self-confidence in diagnosing and managing FH.
Preintervention discussion schedule for general practice staff
Step 2: electronic screening
Using a validated electronic data extraction method (TARB-Ex),19 electronic medical records of all patients seen at the practice in the last 2 years with a recorded blood LDL-c will be screened for potential FH risk. TARB-Ex has an inbuilt standardising algorithm to nominally adjust cholesterol levels (see online supplementary information 1) of patients prescribed statins within 1 week to 6 months of the date of highest recorded cholesterol measurement. Potential FH risk is established using a modified DLCNC score2 and a list of at-risk patients generated. A list of all patients with an existing diagnosis of FH will also be extracted to establish baseline numbers of known FH cases at the practice.
Supplementary file 1
Step 3: medical record review
As part of the intervention and in addition to the previously delivered information sessions, GP/PNs will receive enhanced guidance and training on the assessment and management of patients with FH via one-on-one and face-to-face practice meetings and referred to online information sessions.24 25 In consultation with the clinical investigative team (TB, AV and GFW), the GP/PN team will then review the clinical records of potential FH risk patients (those with TARB-Ex DLCNC scores ≥5) identified in step 2 to determine if they are considered at high risk of FH and require clinical assessment. Other potential causes of hypercholesterolaemia (cholestasis, nephrotic syndrome, steroid use, hypothyroidism) also need to be considered in the decision. Clinical records of existing patients with FH identified in step 2 will also be reviewed by the GP/PN team in consultation with the clinical investigative team (TB, AV and GFW) to determine whether current treatment is optimal and/or if clinical reassessment is required.
Step 4: clinical assessment
The GP/PN will then recall the patients identified at risk of FH and likely to benefit from clinical assessment in order to update family and personal history, exclude possible secondary causes and assess adherence to cholesterol-lowering medications (if necessary). They will also undertake clinical examination to check for corneal arcus and tendon xanthomata. Information on family history and the importance of undertaking cascade testing of family members will also be flagged by the GP/PN. The diagnosis of FH will be based on the DLCNC score at clinical examination, and individuals with scores ≥6 will be regarded as having clinical FH. Consultation with the clinical investigative team (TB, AV and GFW) will be encouraged in confirming FH diagnoses. All newly diagnosed index cases will be consented into the study by the GP/PN. Patients with existing FH identified in step 1 will also be recalled and consented into the study. The general practice team will be encouraged to work with the Australian FH Registry26 27 to record information on newly diagnosed patients as well as any existing patients with FH who are currently unknown to the FH Registry.
Step 5: management and cascade testing
Management of patients is based on consensus opinion6 28 and is outlined in figure 2. Patients assessed as low-to-intermediate complexity FH will be managed by the GP/PN with additional specialist support if required. High-complexity cases will receive additional support from the specialist. Patients’ lipid levels and cholesterol-lowering medication will be monitored (suggest every 3–6 months) as will other lifestyle factors. The GP/PN team will be responsible for cascade testing of FH relatives. Though we have provided guidelines,28 this part of routine clinical practice is likely to prove challenging for research practices initially both from the GP/PN and patients’ perspective. Additional support will be provided as required. In addition, GPs will be encouraged to consent into the study any new patients with FH they may identify as part of usual care.
Suggested management plan derived from: (1) ref 28. (2) Brett, Watts, Garton-Smith, Bell, Vickery, Pang and Arnold-Reed (2015): Familial hypercholesterolaemia: challenges in primary care. Medicine Today 2015; 16(8): 20–26. CVD, cardiovascular disease; FH, familial hypercholesterolaemia; GP, general practitioner; LDL-c, low-density lipoprotein cholesterol; PN, practice nurse.
Step 6: study follow-up
At the end of 12 months, TARB-Ex will be rerun to identify the total number of new FH cases diagnosed over the study period (including those identified through TARB-Ex, cascade-tested cases and new cases to the practice). Clinical records of all consented patients with FH detected will be reviewed and data on treatment and most recent LDL-c levels will be extracted from medical records to evaluate any change in LDL-c over the study period. All patients diagnosed with FH will be encouraged to have their information included in the Australian FH Registry.
Step 7: follow-up qualitative interviews
Patient, GP, PN and practice staff perceptions of the enablers and barriers and suggested improvements to using the approach in addressing FH will be examined (tables 3 and 4). This will be achieved using qualitative methodology which includes a triangulation of data collection methods including both semistructured interviews (face-to-face and telephone) and focus groups (poststudy community feedback sessions). The exact method will be guided by the availability and preferences of the participants and it is acknowledged that the different methodologies used will elicit different information. For example, the preference potentially from a time perspective may be for GPs to complete a telephone interview. Conversely, in order to achieve a balance of individual perspectives, focus groups will be conducted among patient groups across practices. Specialist views on successes and failures on the overall process will also be elicited (table 5).
Discussion schedule for general practice staff postintervention
Discussion schedule for patients postintervention
Discussion schedule for specialist staff postintervention
Step 8: economic assessment
The economic implications of implementing the MoC will be assessed. The costs of developing the TARB-Ex screening protocol have been identified in a previous paper.19 In addition, labour costs of implementing the MoC including costs of training GPs and PNs in screening patients will be tracked at each practice. Furthermore, administrative data from each primary care practice will be sourced to identify the costs of managing patients with FH after they have been identified. Where FH cases are referred to specialist care, administrative data from the tertiary centres they were referred to will be sourced to identify the specialist costs associated with managing these patients. These costs can then be averaged to provide a mean cost per FH case.
Sample size calculation
Our sample size calculation is based on the number of patients needed to detect a statistically significant reduction in LDL-c at 12-month follow-up. Based on clinical audit data from three WA practices, the expected reduction in LDL-c for patients with FH receiving intensive statin treatment over 12 months is 2.83±1.15 mmol/L (n=13; Vickery, personal communication) versus 0.35 mmol/L in the control arm based on data from our previous study based on usual care.29 This is an expected mean difference of 2.48 mmol/L. If we assume a more conservative mean difference of 2.0 mmol/L, we will need 41 patients with clinically diagnosed FH to adequately power our study at a 0.80 level with a type I error probability of 0.05 (two tailed).
Data from our pilot study19 suggest the prevalence of FH in general practice is 1:412. Therefore, to detect 41 cases of clinically diagnosed FH, we will need to screen 16 892 patients.
To adjust for variation across five practices (clustering effect), we have applied a design effect inflation adjustment of 1.6 based on an assumed intraclass correlation coefficient for process variation in primary care of 0.15.30 This estimates to a required total sample size of at least 27 027.
Pilot data19 show that we are likely to have on average 5–8000 patients per practice who have been seen in the last 2 years. Hence, the study is adequately powered to achieve the primary outcome.
Expected extent of cascade testing to inform proposed sample size
Recruitment from five practices will generate an expected 66 index cases. It is expected that up to 50% of relatives will be diagnosed with FH,31 potentially providing between 66 and 125 FH cases. However, it is acknowledged that not all relatives will be followed up as they may not be patients of participating practices or may decline follow-up.
Outcome measures
Quantitative data
number of known index cases of FH and new cases identified through TARB-Ex;
highest LDL-c measure ever;
LDL-c measure closest to baseline;
LDL-c measure closest to 12-month follow-up;
statin type, dose and length of time prescribed over study period;
other CVD risk factors present;
attempts recorded in notes to contact family members;
number of family members with existing or new FH diagnosis or contacted.
Cost data
implementation and extraction of patient data using TARB-Ex;
GP/PN time involved in manual screening of patient records;
personnel time/resources involved in recall of patients;
GP/PN time (from billing schedules) involved in clinical screening of patient records;
patient management costs based on number of visits, length of visits for GP practice and specialist tertiary referrals over the course of the study;
number and type of prescriptions issued.
Qualitative data
Patient and clinical staff views will be assessed using items listed in tables 2–5.
Data management
Data will be deidentified prior to records leaving the GP practice. Only deidentified data will be used by researchers. Audio recordings collected as part of the interviews and focus groups will be kept in secure, locked cabinets. No identifying patient information (eg, names, age) will be recorded during the interviews. Research data will be stored in a password-protected database on a secure server. This network to which the server belongs is protected by multiple firewalls to restrict outside, unauthorised access. Only aggregated, non-identifiable data will be produced in hard copy and this will be maintained in a ‘restricted-access’ locked cabinet located in a lockable office when not in direct use. Only specified research personnel will be permitted access to the data. As this is a clinical study, data will be retained for 15 years after publication.
Data analysis
Statistical methods
Analyses will be conducted using Stata V.13. Descriptive statistics will be used to outline the number of patients at each point of the study (screened, at risk, followed up and clinically diagnosed with FH). At 12 months, the number of new index cases and number of new family cases detected will be reported. FH prevalence will be calculated as the total number of index cases as a proportion of all active patients. The FH detection rate will be calculated as the number of new, clinically identified index cases at 12 months as a proportion of the number of at-risk patients identified through TARB-Ex at baseline. Change in LDL-c will be examined using multilevel mixed-effects modelling. LDL-c level at 12 months will be the dependent variable. Random effects included: GP (cluster effects) and time (repeated observations on the same individual). Analysis will also be adjusted for sex and age.
Cost efficiency analysis
Costs collated from the start of the study to the endpoint at 12 months will be compared against historical costs sourced from tertiary centres for treating and managing FH cases. This will provide an indication of expenditures or savings from adopting the primary care-based MoC.
Qualitative analysis
Qualitative analysis of interview data will be thematically analysed using QSR NVivo V.10. Thematic analysis allows for identification, analysis and detailed reporting of repeated themes within data.32 Interviews will be audio-recorded and transcribed verbatim, then coded for key themes (words, sentences or phrases). An experienced qualitative researcher will be guiding the analysis. Two researchers will be analysing the data to increase reliability of coding and to enable a consensus regarding key themes and subthemes that emerge.
Ethics and dissemination
Ethics approval
The University of Notre Dame Australia Human Research Ethics Committee Protocol ID: 0 16 067F. Australian New Zealand Clinical Trials Registry ID: 12616000630415.
Patient consent
The protocol involves the retrospective screening of patient records, clinical management of patients with FH and eliciting patient and clinic staff views on the method of care. It would not be practical to obtain consent from all patients whose results are to be electronically screened (steps 2 and 6). For patients screened, only deidentified aggregate data will be used for research analysis. For patients who are confirmed as having FH, management will be under GP care but patient-specific information (steps 4 and 5) will be required for analysis. Consent to use this information will be obtained by the GP/PN. At consultation (step 4), the GP will inform the patient of the study. If the patient is confirmed as having FH, the GP will provide the patient with an information sheet and consent form. If the patients consent to be part of the study they will be considered index patients. If they do not consent they will not be followed up for research purposes, but this will not affect treatment received from their GP using study protocol. Step 4 will also provide the opportunity to initiate discussions with those who have consented to be part of the study for potential cascade testing of other family members should the individual patient be considered at high risk for FH. If the family member(s) identified are part of the same practice or seen by the same GP, they can also be considered as an index case and consented to be part of the study. If the new family member(s) identified are not part of the same practice/GP then they will be contacted by their GP (protocol outlined in Brett et al 28) but will not be consented to the study.
GP, PN, practice managers and specialist consent
Steps 1 and 7 involve patient, GP, PN, practice staff and specialist staff perceptions of the success and/or barriers and suggested improvements to the approach. Informed consent will be obtained. The researcher will explain to the participants the purpose of the study and provide them with an information sheet. Consent will be written or verbally recorded.
Confidentiality
Patient medical records will not be removed from the clinic for research purposes. No identifying patient information (eg, names, age) will be recorded during the interviews.
Dissemination
Information will be disseminated via research seminars, conference presentations, journal articles and media releases. Community dissemination is envisaged through community advocacy groups involved in the community engagement at study set-up.
Access to data
The funders shall always be allowed to request full access to all deidentified data collected in the study in a deidentified form as long as this does not violate Australian legal requirements, including, but not limited to, the Privacy Act 1988,33 for research, educational and registration purposes after the conclusion of the study. The funders may request an independent third party to undertake a second evaluation and analysis of data collected in the study.
Ethical considerations
From a patient management perspective, obtaining permission for cascade testing relatives will need to be considered and the protocol outlines suggestions of how this could be done.28 Relatives who are not patients of the practice should be advised by the index case of their potential high FH risk and encouraged to have their LDL-c levels assessed at their local practice. If there is any uncertainty, Commonwealth and State Legislation, National Health and Medical Research Council guidelines and local health service protocols about disclosure of medical information without consent should be followed.
Acknowledgments
We thank the staff and patients at the participating general practices for their assistance in the study.
References
Footnotes
Contributors DEAR, TB and GFW developed the study protocol. DEAR, TB, GFW, JP, LT, IL, JGS, DB and AV contributed to review of the protocol. LT, DEAR and TB developed patient record extraction methods. CB, TG, DEAR and TB developed qualitative data collection questionnaires. LT and MB provided biostatistical input. IL provided input regarding health economic analysis. TG, DEAR and TB drafted and oversaw submission of the application for ethics approval and clinical trials registration. All authors contributed to the review and revision of the manuscript for submission.
Funding The study is funded by Sanofi-Aventis Australia Pty Ltd under an Investigator Sponsored Study Agreement (Study number DIREGL07823). The General Practice and Primary Health Care Research Unit is partly funded by the University of Notre Dame Collaborative Research Network (CRN) programme from the Australian Government. The funders had no input in the design of the study and will have no input in the analysis and interpretation of results.
Competing interests DEAR, AV, JGS, DB, JP, CB, IL and MB report grants from Sanofi-Aventis Australia Pty Ltd during the conduct of the study. TB reports grants from Sanofi-Aventis Australia Pty Ltd during the conduct of the study; personal fees and non-financial support from Sanofi-Aventis Australia Pty Ltd, outside the submitted work. GFW reports grants from Sanofi /Regeneron during the conduct of the study; grants and personal fees from Sanofi/Regeneron, grants and personal fees from Amgen, personal fees from Gemphire, personal fees from Kowa, outside the submitted work. LT reports grants and other from Sanofi-Aventis Australia Pty Ltd during the conduct of the study. TG reports other from Sanofi-Aventis Australia Pty Ltd during the conduct of the study.
Ethics approval The University of Notre Dame Australia Human Research Ethics Committee Protocol ID: 016067F. Australian New Zealand Clinical Trials Registry (ANZCTR) ID: 12616000630415.
Provenance and peer review Not commissioned; externally peer reviewed.