Hearts and Minds- Information to Manage Metabolic Syndrome in Mental Illness

• Abstract
• 1. Introduction
• 2. Pharmacovigilance Background
• 3. Informatics Strategies: General
• 4. Informatics Strategies: Empowerment
• 5. Summary of Project Aims
• 6. Design of the Proposed Database and Preliminary Hypothesis Testing
• 7. Methodological Considerations
• 8. Conclusion
• 9. Acknowledgements
• 10. References

Abstract
Patients treated chronically with antipsychotic medication are at increased risk of metabolic syndrome, cardiovascular disease and premature death. While common and readily identifiable, this problem requires multidisciplinary input and is yet to be effectively managed by health services around the world, including New Zealand’s. As part of developing a New Zealand national programme to address this issue, we consider an informatics strategy to provide essential information to clinicians, researchers, and patients themselves.

1. Introduction
People with severe mental illness, notably schizophrenia and bipolar affective disorder, have increased mortality rates compared with the general population.^{[1, 2]} Much of this excess mortality, particularly among patients treated with antipsychotic drugs, is attributable to cardiovascular disease. It is now understood that metabolic syndrome (central obesity with various combinations of impaired glycaemic control, dyslipidaemia, and hypertension) plays a key role in the genesis of cardiovascular disease, and is aggravated by treatment with antipsychotic drugs.^{[3, 4]} Dyslipidaemia, for example, may lead to diabetes,^[5] and both together interact to increase risk of heart attacks and stroke.

Individuals with chronic mental illness are at particular risk for developing metabolic syndrome because of their tendency to poor diet and/or lack of exercise. Thus, the combined effects of lifestyle and medication mean that people with serious mental illness face the added burden of poor physical health and premature death. The situation for Maori appears, if anything, to be even worse than for European New Zealanders,^[6] as their burden of cardiovascular disease and premature mortality is excessive even before the aggravating effects of antipsychotic medication are considered.^{[7, 8]} The serious threat to public health posed by metabolic syndrome, while recognised, is yet to be effectively managed by health services around the world, including New Zealand’s. The importance of addressing this problem has been emphasised in health policy both in New Zealand,^[6,9] and overseas. ^[4]

The reasons for inadequate management of metabolic syndrome in New Zealand and overseas are multiple, complex, and incompletely understood. Relevant issues include:

• Patient factors:
  • Lack of insight, awareness, or concern
  • Poor concordance with medical and other recommendations regarding mental and physical health
  • Anxiety or lack of confidence regarding doctors, hospitals, or medications.
• Health professional factors:
  • Tendency to blame patients for obesity, lack of fitness or poor compliance
  • Deficient knowledge, skills or experience in caring for this population
  • Pessimism regarding outcome of serious mental illness, metabolic syndrome, or both
  • Poor morale, tiredness, "burnout".
• Institutional and cultural factors:
  • Fragmentation of inpatient and outpatient services
  • Poor communication between primary and secondary care
  • Inadequate resources available for quality multidisciplinary care
  • Pharmacocentric psychiatry education, training, and CPD.

Many of the evident factors reflect the inconsistent and often disorganised state of mental health information. In developing this programme, our overarching philosophy is that effective management of metabolic syndrome will require an effective informatics strategy. As described below, a multidisciplinary programme to address this problem depends on a robust and convenient system of information availability and exchange.

A team of clinician-researchers based in Hamilton (Waikato Clinical School and Waikato District Health Board), together with health informatics experts (Universities of Auckland and Otago), are developing a programme to enable clinicians to effectively monitor and manage metabolic syndrome among antipsychotic-treated patients. Because at least 2.5 percent of New Zealand adults are dispensed antipsychotic medication in a given year (PHARMAC), up to 80,000 are likely to be taking these medications continuously. Some degree of metabolic syndrome is likely in at least one-half of these, with frank complications (diabetes mellitus, hypertension, dyslipidaemia) in one-quarter or more. Thus around 20,000 New Zealand adults are likely to have serious health impairment, with increased risk of premature death, caused or aggravated by antipsychotic treatment. The need for a register of such individuals is compelling, given well recognised difficulties in organising screening, intervention and follow up. Particular consideration of Maori patients is appropriate, given evidence of their increased morbidity and mortality from metabolic syndrome.^{[6, 8]} The same is likely to apply to patients of Pacific Island origin.

The project described here includes a pilot study of methodology for both monitoring and intervention in metabolic syndrome associated with antipsychotic drugs. Evaluation of the pilot will be essential to refine intended nationwide implementation of agreed guidelines (see below). Because of the apparent importance of ethnicity, an "oversampling" strategy may be necessary to ensure comparable numbers of Maori and European New Zealanders in the study cohort ^[10] and thereby optimise statistical power. The proposed project is thus important as a contribution toward management of an urgent and extensive unmet clinical need in New Zealand; it may also have value as a more general model system to develop distributed information access to clinicians of various disciplines, as well as to patients, researchers and administrators.

2. Pharmacovigilance Background
The Intensive Medicines Monitoring Programme (IMMP) is part of the New Zealand Pharmacovigilance Centre, based at the University of Otago (http://carm.otago.ac.nz/). The IMMP collects nationwide prescription data for selected medicines to establish cohorts of patients who are followed up via prescribers and also by record linkage to New Zealand Health Information Service (NZHIS) databases which contain, eg, data on hospital admissions and deaths. The IMMP, funded by the New Zealand Ministry of Health, thus serves to identify adverse clinical events associated with the medicines in question, and is among the most effective pharmacovigilance agencies in the world. However, capture of most prescription and outcome data is on paper, with reporting delays typically of several months. Among the advantages of "real time" electronic capture of prescription/exposure data would be the possibility to expand the database as a tool for general clinical use. We have thus agreed to work with the IMMP to develop the database for the purpose of monitoring and managing metabolic syndrome. The IMMP database provides a starting point of particular advantage since each of the main four second generation antipsychotic drugs (SGAs) used in New Zealand have been monitored for some years. Among antipsychotics, these drugs (clozapine, olanzapine, quetiapine, risperidone) are the most consistently associated with the induction of weight gain and metabolic problems.

3. Informatics Strategies: General
Programmes to improve pharmacovigilance databases such as IMMP should focus on the twin goals of developing health intelligence and empowering patients. These goals reflect our overarching aim to get the maximum leverage from data collected to understand and exploit the underlying health knowledge.

Health intelligence can be defined as "tools, methods, technologies and processes to transform data into knowledge, and to apply such knowledge". Using such knowledge for clinical decision support can improve the quality of care by optimising health benefits and reducing risk of harms, eg, by facilitating clinician adherence to evidence-based best practice. Health intelligence is more than the ability to visualise data with data mining; it uses knowledge to predict, optimise and adapt. Figure 1 illustrates how data could progress to knowledge for making decisions, based on prediction, optimisation and adaptability. Data are collected in the form of bits, numbers, symbols, and objects. Information is organised data, which are pre-processed, cleaned, arranged into structures, and stripped of redundancy. Knowledge is integrated information, which includes facts and relationships that have been perceived, discovered, or learned.

Health intelligence could help answer:

• What is the best decision right now?
• What is likely to happen in the future?
• What should be done if X occurs?

  Figure 1: Data to Knowledge (from reference 11)

  

  Drug treatment knowledge can be developed from clinician case reports, and from data-mining techniques coupled with traditional statistical methods. Data-mining will include natural language and artificial intelligence techniques.

  
  4. Informatics Strategies: Empowerment
  The patient voice is getting louder; increasingly patients and their families/whanau want shared decision making. The trend with on-line health and electronic health records reflects the increasing role of the patient to be an active partner in care planning and taking greater responsibility for their own health care. Online disease self-management can be an effective delivery method for teaching patients the skills and self-confidence they need to take charge of their disease care.^[12] This is now further supported by the growth of Web 2.0 and on-line "social spaces". The industry analysis company Gartner Group predicts that Web 2.0 will ". . . impact on a broad range of traditional enterprises".^[13]

  Present solutions for the retrieval and representation of medical information from online sources are not very satisfying.^[14] Either the retrieval process lacks accuracy or the representation does not support updating and maintenance of the information. In the Adverse Drug Reaction (ADR) domain, consumer empowerment has very recently been improved by the introduction of DoubleCheckMD onstage (http://www.doublecheckmd.com/) and PharmaSurveyor. DoubleCheckMD onstage is operable in a Beta format. PharmaSurveyor is due for release in 2008.^[15] Both products allow individuals (and doctors) to mine huge amounts of information to assess their own drug combinations.

  DoubleCheckMD onstage is claimed to be the first online medical search technology that empowers users to easily find accessible, accurate and information on drug interactions and side effects. Our project Hearts and Minds aims to be similarly empowering. Natural language, statistical analysis and artificial intelligence techniques will be used to create knowledge from the data. In addition Web 2.0 technologies could be used to form on-line communities that can share experiences of drug-related problems, eg, side effects, to further the empowerment and knowledge creation aims of Hearts and Minds.

  Both knowledge development and doctor-patient empowerment require an extension of the functionality of IMMP and current mental health databases. There is arguably a need to integrate legacy systems and analyse information in a more holistic manner with semantic web type ontologies. Anticipated problems include what to do if the system puts up several solutions. In addition, some clinicians may not be proficient in using new technologies, or in applying evidence-based practice, and thus unable to make effective use of the solutions.

  
  5. Summary of Project Aims
  1. To develop a national programme to:
     1. identify all patients dispensed antipsychotic medications in New Zealand with real-time reporting from all authorised pharmacies and hospitals; and
     2. capture and organise relevant demographic, pharmaceutical, anthropometric, laboratory and other relevant data for such patients.
  2. To facilitate the appraisal of the physical health of antipsychotic-treated patients, specifically considering evidence for metabolic syndrome.
  3. To analyse risk factors, in terms of drug history, laboratory and anthropometric measures, for metabolic syndrome.
  4. To specifically assess the role of ethnicity (in the first instance, Maori vs European New Zealanders) in the genesis of metabolic syndrome.
  5. To develop an integrated set of algorithms to guide monitoring, treatment and follow-up of metabolic syndrome in this cohort (see figure 2).
  6. To facilitate collegial communication and referral, in order to enhance clinical care and outcomes.
  7. To provide information for research regarding:
     1. detection of metabolic syndrome and its complications;
     2. guidance, monitoring and follow-up of treatment; and
     3. genetic and other investigations relevant to incidence, management and outcome of metabolic syndrome.

  Figure 2: Flowchart for monitoring and management of metabolic syndrome

  

  
  6. Design of the Proposed Database and Preliminary Hypothesis Testing

  1. A pilot study of a defined cohort (300 chronic patients in each of two urban areas: Hamilton and Dunedin) are proposed to guide onward national programme development, in particular:
     1. analysis of record retrieval and data collection will yield insights regarding the availability of relevant information, its location and means of optimal retrieval, in long term patients treated with antipsychotics.
     2. collection of face-to-face data will allow completion of the dataset, and validation of available records, with regard to key information: notably current drug treatment, anthropometry, and ethnicity.
  2. Relevant information (drug prescription history, physical health measures) will be used systematically to predict metabolic syndrome risk in long term patients.
  3. Specific comparison of Maori and European New Zealand patients will allow statistical estimation of the contribution of ethnicity to metabolic syndrome risk in long term patients.^[16]
  4. DNA sampling, with detection of candidate genes involved in obesity, diabetes, and cardiovascular disease, will add to prediction of:
     1. metabolic syndrome and its consequences
     2. response to intervention (see below).
  5. Development and implementation of an appropriate user interface will allow access to relevant information for clinical and other staff with a need to know, and others:
     1. doctors: psychiatrists, GPs, endocrinologists and other specialist physicians
     2. nurses: inpatient and community mental health teams
     3. pharmacists: hospital and community
     4. dieticians or other staff involved in promoting healthy eating
     5. physical education, occupational therapy, or other staff involved in promoting physical activity
     6. researchers
     7. administrators
     8. patients and, where appropriate, their families.
  6. Factors predicting metabolic syndrome and its consequences will also predict usefulness of various treatments for dyslipidaemia, diabetes, and hypertension, including:
     1. pharmacological (eg, statins, hypoglycaemics, thiazides)
     2. dietary manoeuvres
     3. physical activity programmes.
  7. Management algorithms can be effectively built into the information system to:
     1. provisionally assign patients to clinical categories, and thereby provide guidance to clinicians regarding assessment and treatment
     2. provide prompts for medication review and prescription renewal
     3. formulate and transmit referrals
     4. inform relevant staff of changes to drug and other treatment
     5. prompt needed follow up regarding clinical and laboratory investigations.
  
  7. Methodological Considerations
  1. Sampling of chronic cohort will be used to define target sample (in each centre, 150 patients in each of two groups self-identified as Maori and European New Zealanders; this is expected to be straightforward in Hamilton but will require oversampling of Maori in Dunedin.^[16] Ethnicity provisionally to be determined as per New Zealand Health Information Service protocol (http://www.nzhis.govt.nz/documentation/ethnicity/ethnicity-05.html#question).
  2. Retrieval and entry of current data will include NHI, demographics, psychiatric and other drug treatment, anthropometric and laboratory measures according to national guidelines,^[17] and documented physical illness.

  
  8. Conclusion
  There is now compelling evidence that health informatics will be a powerful vehicle for effecting change leading to better mental and physical health outcomes in chronic illness. In particular, a comprehensive informatics strategy promises to provide clinicians with needed data and decision support to effectively monitor and manage the scourge of metabolic syndrome among patients requiring treatment with antipsychotic medications.

  
  9. Acknowledgements
  The authors thank Jim Warren (Informatics, The University of Auckland), Mira Harrison-Woolrych (IMMP, Dunedin) and Michael Mair (Ophthalmologist, Timaru) for discussions. Planned staffing of the overall project includes the following:

  Hamilton Clinical Staff

  1. Psychiatrists: David Menkes (lead, and pharmacologist), Adrian Leathart, Andrew Darby, Rajiv Singh (clinical director)
  2. Pharmacists: Neville Puckey, Guna Kanniah, Erica Amon (PHO)
  3. Endocrinologist: Ross Lawrenson
  4. Nurses: Kathryn Kempson, Theresa Carroll, Brian Thomas, Luren Reddy
  5. Occupational Therapist: Andrew Parkin
  6. Physical educationalist: Stephanie McClennan and Jen Riley (Sport Waikato)
  7. Dietician: Viv Dykes

  Dunedin Clinical Staff

  8. Psychiatrists: Richard Mullen, Chris Gale, James Knight (clinical director)
  9. Pharmacist: Lucy Broughton, David Woods (also BPAC), others to be recruited
  10. Endocrinologist: Jim Mann
  11. Nurse: Bernadette Forde (local lead), others to be recruited
  12. Occupational Therapist: to be recruited
  13. Physiotherapist: to be recruited

  Other Staff

  14. Pharmacovigilance: Mira Harrison-Woolrych, Janelle Ashton and David Clark (IMMP, University of Otago, Dunedin)
  15. Genetics: Martin Kennedy, et al (Gene Structure and Function Laboratory, Christ-church School of Medicine)
  16. Informatics and Decision Support: Martin Orr (Waitemata DHB and University of Auckland), Jim Warren (The University of Auckland), David Ireland and Scott Elli-ott (Waikato DHB), Alec Holt (University of Otago), Sue McArthur (Smartmed), Michael Mair (South Canterbury DHB and New Zealand HL7 User Group)
  17. Statistician: Gaelle Dutu (Waikato Clinical School)

  Advisory Group

  18. Psychiatrists: Rees Tapsell and Lyndy Matthews (Auckland, Hamilton, and RANZCP), Peter Ellis (Wellington and RANZCP), David Healy (Bangor, UK)
  19. General Practitioner: Murray Tilyard (University of Otago, and BPAC)
  20. Pharmacist: Keith Crump (Waitemata DHB; ProCare Education), Jeff Harrison (The University of Auckland)

  
  10. References
  1. Fichter MM, Rehm J, Elton M, Dilling H, Achatz F. Mortality risk and mental disorders: Longitudinal results from the Upper Bavarian study. Psychological Medicine 1995;25(2):297-307.
  2. Osborn DPJ, Levy G, Nazareth I, Petersen I, Islam A, King MB. Relative risk of cardiovascular and cancer mortality in people with severe mental illness from the United Kingdom’s General Practice Research Database. Archives of General Psychiatry 2007;64(2):242-249.
  3. Nasrallah HA, Meyer JM, Goff DC, McEvoy JP, Davis SM, Stroup TS, et al. Low rates of treatment for hypertension, dyslipidemia and diabetes in schizophrenia: data from the CATIE schizophrenia trial sample at baseline. Schizophrenia Research 2006;86(1-3):15-22.
  4. Newcomer JW. Antipsychotic medications: Metabolic and cardiovascular risk. Journal of Clinical Psychiatry 2007;68(Suppl. 4):8-13.
  5. Barner JC, Worchel J, Yang M. Frequency of new-onset diabetes mellitus and use of antipsychotic drugs among Central Texas veterans. Pharmacotherapy 2004;24(11):1529-38.
  6. Poa NR, Edgar PF. Insulin Resistance Is Associated With Hypercortisolemia in Polynesian Patients Treated With Antipsychotic Medication. Diabetes Care 2007;30:1425-29.
  7. McPherson KM, Harwood M, McNaughton HK. Ethnicity, equity and quality: lessons from New Zealand (Nga matawaka, nga ahua tika me nga painga: nga akoranga no Aotearoa). [Editorial]. Quality & Safety in Health Care 2003;12:237-8.
  8. Simmons D, Thompson CF. Prevalence of the Metabolic Syndrome Among Adult New Zealanders of Polynesian and European Descent. [Report]. Diabetes Care 2004;27:3002-4.
  9. Te Kokiri: The Mental Health and Addiction Action Plan 2006-2015. Wellington: Ministry of Health; 2006.
  10. Mana Whakamarama -- Equal Explanatory Power: Maori and non-Maori sample size in national health surveys. Wellington: Wellington School of Medicine; 2002.
  11. Michalewicz Z, Schmidt M, Michalewicz M, Chiriac C. 2006, Adaptive Business Intelligence. Springer, 246p
  12. Greenhut JH. (2007) Patients Learn Chronic Disease Self-Management Online. Medscape Public Health & Prevention. http://www.medscape.com/viewarticle/558483 retrieved 16th July 2007.
  13. Nadler L. (2006) Gartner Highlights Seven Core Benefits of Web 2.0 for Traditional Industries. http://www.gartner.com/it/page.jsp?id=499154 retrieved 16th July 2007
  14. Mann G, Birkmann C, Schmidt T, Schaeffler V. (1999) COM3/369: Knowledge-based Information Systems: A new approach for the representation and retrieval of medical information. J Med Internet Res 1999;1(suppl1):e16 http://www.jmir.org/1999/suppl1/e16/
  15. Von Schweber, E, 2007, Semantic Web Use Cases and Case Studies, Case Study: Composing a Safer Drug Regimen for each Patient with Semantic Web Technologies. PharmaSURVEYOR Inc. http://www.w3.org/2001/sw/sweo/public/UseCases/PharmaSurveyor/
  16. Mana Whakamarama -- Equal Explanatory Power: Maori and non-Maori sample size in national health surveys. Wellington: Wellington School of Medicine; 2002.
  17. New Zealand Mental Health Metabolic Working Group Initiative. Auckland: Janssen Cilag, Ltd.; 2006.