How to avoid patient nonadherence
Learn more on why patients fail to follow doctors’ recommendations and how to improve patient adherence in eye care.
Patient noncompliance, also called nonadherence, is one of the most common causes of treatment failure in eye care. A patient’s failure to follow the recommendations of his physician shatters quality healthcare outcomes and carries an appreciable economic burden, according to Jay Lytle, OD, FAAO, during a talk at the American Academy of Optometry (AAO) Academy at Home 2020 virtual meeting.
In the session, Lytle defined adherence, explored incidence rates, reviewed financial and medical implications of nonadherence, suggested in-office assessments, and identified adherence killers and promoters.
Defining adherence
Medication adherence is the extent to which a patient adheres to the prescribed dose and interval of his treatment regimen. Patients are considered adherent when the number of doses taken by the patient is 80% or greater than the prescribed doses.1
The first mention of the word “compliance” in medical literature appeared in 1966.2 In 2000, the word, “adherence” became preferred over the word, “compliance” among experts.2 By 2010, many researchers published meta-analyses summarizing hundreds of past articles on nonadherence.1,3-4
Today, nonadherence is one of the most commonly researched topics in medical journals.5-14
Related: Improve medication adherence with technology
Incidence rates
Most studies on nonadherence come from endocrinology and cardiovascular medicine, but the overall nonadherence rate for both acute and chronic care combined, across all areas of medicine, is 40%.15,16 When it comes to chronic asymptomatic conditions, the nonadherence rate jumps to 80% across all areas of medicine.1
When it comes to eye care, optometry and ophthalmology have focused research on adherence only within in the past 20 years. For dry eye patients with topical therapy, the average discontinuation time was 3 months for Restasis (cyclosporin, Allergan) and 1 month for Xiidra (lifitegrast, Novartis).17
For pediatric uveitis patients, where parents are administering the medication for their children, the nonadherence rate is 42%.18 And for glaucoma patients using topical agents, the nonadherence rate is 59%.19
Problem variables
Many of the studies that evaluate nonadherence attempt to identify contributing variables. And in those studies, nonadherence with once-daily dosage is 16% on average.16,17 But when doctors increase that dosage to 3 times a day, nonadherence goes up to 41%.16,17
Problem variables that have been quantified in incidence rates include:
Dosage regimens
Treatment complexity
Poor doctor communication
“The results of multiple different studies show that each additional dose of medication causes a 10% increase of nonadherence,” says Lytle. “These results highlight the importance for ODs to prescribe the lowest therapeutic dose required for a medication.”
Public health impact
The weight of a 40% global nonadherence rate is magnified when up to 33% of Americans are living with one or more chronic diseases that may require treatment, such as high blood pressure, high cholesterol, diabetes, respiratory problems, and arthritis.1 Some 117 million people are living with these conditions.19
In eye care, 2 different studies found that nonadherence with pediatric uveitis resulted in 10 times higher risk for failed treatments.10,18 Another study found that improving glaucoma adherence led to 4 additional years of sight.11
Improving medication adherence may have a greater public impact than the discovery of new medications because many effective medications already exist for most chronic conditions.1
Related: Why noncompliance will ruin your perfect treatment plan
Economic impact
In the United States, almost $300 billion is spent every year on avoidable healthcare costs due to patient nonadherence.1,15
“That’s 10% of all the healthcare expenditures in the United States on unnecessary expenses,” Lytle says. “The United States alone wastes more money not listening to doctors than most countries produce in their entire economies.”
The U.S. nonadherence cost equates to 700,000 emergency room (ER) visits for diabetes alone and it also amounts to 125,000 avoidable deaths each year.1,13,16,20
The effects also take a toll on doctors. One study found that doctors struggle with the negative effects of anger, hopelessness, betrayal, and regret as a result of patient nonadherence.1
Why are patients nonadherent?
Several studies identify 6 reasons for patient non-adherence with 1 major offender. The 6 reasons for patient nonadherence are:
Poor recall or misunderstanding16,21
Poor execution or misapplication16
Poor memory or forgetting16,21
Poor access or lack of finances/transportation1,21
Poor experience or side effects21
Poor buy-in or intentional nonadherence16,21
The primary reason for patient nonadherence, according to global clinical studies, is poor buy-in, which accounts for billions of dollars in wasted emergency visits, years of lost sight, and physician burnout.1,16,21
“The number 1 reason for patient nonadherence is an intentional choice on the part of the patients solely because they don’t believe what their doctors said,” Lytle says. “To make matters worse, due to social desirability bias, intentional nonadherence is often accompanied by concealment because patients don’t want to disappoint their doctors, they feel embarrassed, or they fear their doctors will punish them for not adhering to the prescribed treatment.”
Related: Rethinking once-daily dosing on allergy medications
Tracking adherence
Adherence is tracked 4 ways in clinical studies. One way is health outcomes, which retrospectively analyzes medical records or the number of hospitalizations for a given condition.
“This method is not the best,” Lytle says.
The most accurate method of tracking adherence is biological markers like blood pressure, interocular pressure (IOP), or bloodwork.
“If we could accurately get this information on everyone, it would be the most reliable method,” Lytle says.
The third method of tracking adherence is indirect indicators, such as whether patients went to the pharmacy to refill medications or whether they attended appointments. Lytle says trackers placed on pill or eyedrop bottles are able to tell doctors whether a drop has been released, but such methods have limitations.
“They would be used only in a closed clinical trial; you wouldn’t use them in common practice because they don’t necessarily indicate that an eyedrop made it into a patient’s eye or that a pill made it into a patient’s mouth,” Lytle says.
The fourth tracking method is a subjective report study. This is when patients journal whether they took the medication, or outside observers like family members or a member of the medical team’s staff checks up on the patient to record their own reports on whether the patient followed through. There is a self-reporting scale called the Morisky scale that is a validated screening measure of patient adherence. But, research shows that this fourth method is not very reliable. Patients, caregivers, and even doctors are not reliable determiners of compliance, even when a physical exam is involved.1,16,19
“The impression of the doctor is not necessarily indicative of the patient’s adherence level,” Lytle says.
The same studies of subjective reporting found that age 12 is the approximate age at which adolescents become primarily responsible for administering their own medications.1,16,19
“If you work with kids and you have a child who is 12 or older and you prescribe an eye drop, you should probably spend more time instructing that child rather than speaking to the parent,” Lytle says. “Because it is very possible that 12-year-old might be expected to put in their own eye drops at home.”
In-office adherence assessment
ODs can address nonadherence by first creating an environment in their offices where they can accurately assess adherence. To do this, create a blame-free environment of mutual trust in which patients feel comfortable reporting their true adherence.
Subtle differences in wording when interviewing patients can significantly improve the concept of a blame-free environment. For instance, instead of asking, “Are you taking your eye drops?” say, “How often were you able to get your eye drops in?” It is more effective to use non-accusatory, open-ended questions, Lytle says.
Researchers have found that the person asking the adherence question should be the person on the clinical team who the patient trusts the most.1 That may not necessarily be the technician who started training last week, and it might not even be the patient’s doctor.
When a patient does admit to nonadherence, clinicians should indicate appreciation for the patient being forthcoming and then validate the reason as legitimate and understandable, even if it is not legitimate or understandable, Lytle says.
This could be as simple as, “Thanks for letting me know. That’s understandable.”
Patients tend to tell the truth only when they feel like they won’t be criticized or judged.1
“It is important for ODs to understand that the most accurate adherence assessments are simple and non-threatening,” Lytle says.
Adherence killers
Five adherence killers repeatedly pop up in medical literature, and a simple ABCDE mnemonic can be used to remember them.
ABCDE stands for:
Access barriers: Financial, transportation
Beliefs: Personal and cultural stigmas, low health literacy
Complex treatments: Number of medications, dosages
Depression: Support withdrawal, lack of hope
Elderly: More medications, cognitive decline, physical limitations
Access barriers
When patients cannot afford medications, they won’t be able to take them. Doctors can address access barriers by first familiarizing themselves with the cost of medications before prescribing them. Online resources like GoodRx.com provide some measure of cost analysis.
For patients who can’t make it to the pharmacy because of transportation barriers, doctors or staff members can connect them to services like ride shares.
Beliefs
When it comes to beliefs, some patients have past negative medical experiences that may create barriers for future encounters. For other patients, cultural beliefs may not coincide with a diagnosis or treatment plan. When these obstacles cloud the management process, having open conversations about what options do coincide with their belief systems can create an environment of prescribing with which the patient can agree.
“One of the single greatest killers of patient adherence is a low health literacy level,” Lytle says.
The U.S. Department of Education estimates that 35% of Americans have low health literacy preventing them from being able to read a medicine bottle or a poison warning.22
The U.S. Department of Health and Human Services defines health literacy as “the degree to which individuals have the capacity to obtain, process, and understand basic health information needed to make appropriate health decisions.”23,24
Health literacy is over 8 times lower in individuals who are over age 85.1,16,18 It is also lower in persons whose primary language is not English and those who have been recently diagnosed with a condition.1,16,18
Fostering two-way communication with patients can effectively prevent the negative consequences of low health literacy, Lytle says. Doing so will allow ODs to determine a patient’s understanding level, then provide him with written and verbal education that he can follow.
Complex treatments
The more complex the treatment, the worse the adherence. For every additional dose of daily medication, adherence drops by 10%, Lytle says. Plus, most serious conditions also tend to have complex treatments.
So, one of the most important things that ODs can do is simplify the therapy to the lowest daily dose and the lowest number of total medications possible while still being therapeutic.
Depression
Some experts believe that depression may have a larger effect on nonadherence than any other variable.
“Patients with depression are 2 to 3 times more likely to have poor adherence than the general population because individuals with depression often withdraw from their support networks,” Lytle says.
It is estimated that 63% of patients with depression are nonadherent.1,2
All doctors should prepare a depression discussion so they can identify depression and connect patients with appropriate professionals to treat it.
“If you have someone in your chair, and you have a strong suspicion that she might have depression, you should have the phone number of a clinical therapist or a primary-care doctor who specializes in the field of depression care in your area,” Lytle says.
When depression goes untreated, adherence drops dramatically.1,3,24 Pre-screening tools, including the PHQ9 form, are quick, standardized questionnaires doctors can give to patients on a clipboard in the reception area. These tools can help identify if a patient is suffering from depression.
Once depression is identified, ODs can connect depressed patients with professionals who can appropriately treat them by saying, “Have you felt down or depressed lately?” and “Someone I trust when it comes to depression is Dr. ___, would you be open to us setting up a visit at his office for you?”
Elderly
Patients over age 70 have higher rates of nonadherence, likely due to the quantity of medications that they manage, but they may also have physical or cognitive limitations, Lytle says. It is critical to ask elderly patients, more than any other group of patients, about support networks to ensure their ability to receive the treatment that they need.
Adherence promoters
The 2 factors identified as promoting adherence are social support and effective doctor-patient communication, Lytle says.
Social support means connecting with a patient’s support system. Doing so can significantly improve the patient’s ability to adhere to a treatment plan. Social support includes 3 things:
Medication support
Transportation support
Encouragement support
“The most successful long-term support systems are multi-layered and intentional,” Lytle says.
Medication support
At-home family support increases medication adherence by 3 times.2 When loved ones are present, they can help individuals who might have physical or cognitive limitations or those who are just plain busy.
If, after having a conversation with a patient, a doctor identifies that there is a gap in that patient’s support system, the doctor’s team can help fill that gap with a phone reminder by calendar or text. This can help the patient to take her medication on time.
Transportation support
If patients can’t make it to the pharmacy or doctor’s office, adherence is severely impaired.2 The doctor’s team can connect patients with ride-share services on their phones within minutes. To alleviate the need for a ride, a mail-order pharmacy can also be set up. Telehealth options can also help to maintain follow-up visits.
Encouragement support
Research has found that having someone who simply cares about the patient and offers him encouragement can have a significant impact on treatment adherence, Lytle says. The person who fulfills this role of encouragement doesn’t necessarily have to be a family member. Research found that being involved in enjoyable social activities was enough to meet this criteria.2,7,20
“Being involved in the local bingo club is enough to meet encouragement support criteria,” Lytle says.
For patients who lack encouragement support, brief phone calls with the doctor or a staff member can be an effective replacement to improve treatment adherence.
Communication road-bumps
Research shows that doctors commit 6 cardinal sins during treatment discussions with patients, Lytle says. As a result, when doctors do ask for questions, patients rarely ask them.25 The 6 communication sins doctors most frequently commit are:
Dominate discussions
Do not discuss risks or side effects
Do not mention the medication name
Do not discuss how new meds differ from failed old meds
Do not check patient comprehension
Believe the diagnosis justifies the medication
In order to foster adherence and create an environment of trust, it is important that doctors provide patients with comprehensive overviews of their diagnosis and treatment. Additionally, ODs should review communication road bumps before one-on-one conversations with patients so they develop a plan for how to avoid them.
Good bedside manner
A study from 2009 found that approximately two-thirds of what we say is never spoken but rather non-verbally communicated.26
Effective non-verbal communication, also known as bedside manner, includes 3 different things:2,16
Body language
Facial expressions
Room position
“Having a healthy work environment and healthy staff members leads to healthier patients,” Lytle says, “likely because of nonverbal cues.”
Good body language involves being seated, with shoulders facing toward the patient and both feet on the floor. Leaning slightly inward is another way to convey positive rapport. Additionally, effective facial expressions like periodic eye contact and smiling can communicate friendliness and rapport.
“There should be no barriers between doctors and patients, including computers, charts, or clipboards,” Lytle says, “and in most Western countries, maintaining a distance of 6 to 8 feet apart is appropriate.”
Verbal communication
When it is finally time to talk, ODs should deliver a treatment plan. Effective verbal communication during this conversation can help improve patient adherence.
Effective verbal communication includes 8 different factors, the first of which is most important. They are:1-4,8,9,16,21,27-29
Reach concordance: Mutual collaboration
Build trust: Positive rapport over time
Choose vocabulary: Tailor to the patient’s health literacy level
Check comprehension: Pause to assess patient comprehension
Start the treatment: Make a definitive diagnosis and start the appropriate treatment
Review the plan: Review the treatment plan at each visit and provide recordings, if desired
Schedule follow-ups: Schedule follow-ups and organize appointment reminders
Provide resources: Provide appropriate educational resources and point patients toward helpful outside sources
Reach concordance
“If you remember one thing from everything mentioned, it should be this,” Dr. Lytle says. “Concordance is a prescribing process where the patient and physician discuss their beliefs about the treatment in an open, mutual negotiation.”
To reach concordance, start by eliciting the patient’s views by saying, “Have you tried anything in the past for this?” Patients are usually hesitant to reveal past remedies unless the doctor opens up the discussion, Lytle says.
Then, explore the patient’s views with a statement such as, “There are a few medications I have had success with. Would you like to talk about our options?”
“Most doctors don’t talk about side effects, dosing, timing, or pricing—but should,” Lytle says.
ODs should open two-way discussions about the risks and benefits of treatments in order to learn the patient’s intentions. Additionally, they should involve the patient in the decision-making process by saying something like, “Is that something you would be comfortable trying?” Doing so empowers patients to become autonomous agents of their own well-being.
Build trust
Building trust is especially important in the prescribing landscape because up to 55% of patients believe that doctors are paid by pharmaceutical companies and have a vested interest in selling medications.1 This belief is highest in young and wealthy patients.
Fostering an environment of mutual trust increases adherence by 3 times.16 When a patient believes that his doctor is neither competent nor caring, the result is distrust. When a patient believes that his doctor is very smart but doesn’t care about him, he might respect his doctor but he will never trust him. When a patient believes that his doctor really cares about him but doesn’t know what he’s talking about, he might like the doctor, but he won’t trust him. True trust begins only when a patient believes that his doctor is both competent and cares about him as a person.
Based on concepts developed by John Spence, business consultant, and Susan Fiske, department of psychology, Princeton University
“It turns out mama was right,” Lytle says, “People don’t care how much we know until they know how much we care.”
Building trust can be achieved in 3 different ways:
Familiarity
Empathy
The golden moment
When doctors first meet patients, researchers recommend doctors give them undivided, uninterrupted attention for a certain length of time, Lytle says. Sometimes, it can be less than a minute for the golden moment to occur, but that is enough to set the tone for all future encounters with that patient. A patient’s very first impression of her doctor should be, “Wow, my doctor really listens closely.”
The second way to achieve trust is with time and familiarity.
“This one is not as easy to achieve quickly,” Lytle says.
Time and familiarity are one path to doctor-patient trust. More quality time and cumulative visits spent with a doctor result in higher rapport and trust. Therefore, new patients may require extra communication to make up for the lack of familiarity.
The third way to achieve trust is with empathy. Empathy here refers to when the patient believes that her doctor can understand her perspective, Lytle says. ODs can express empathy by identifying the patient’s emotions, exploring the reasons for her emotions, and validating her emotions.
The last 6 factors of effective communication are important, but not as important as reaching concordance and building trust, Lytle says.
According to Lytle, most studies have found that doctor-patient trust is not dependent on age, race, or ethnicity. Doctors who provide quality communication can earn the trust of their patients and connect with them regardless of their differences with that patient.
One study from the Netherlands showed that patients cared more about a doctor’s compassion than their culture.16
“The voice of research says that it isn’t the color of our skin, the accent in our voices, the year of our birth, or what holidays we celebrate in December that patients care about,” Lytle says. “Patients care about being treated like a person, and that is both powerful and scary at the same time because it means we have no excuses. There are no inherent roadblocks for us to be successful in our communication.”
In conclusion, ODs may not be able to control some things in their offices, but they can control how they communicate with their patients. Quality communication leads to quality care and a decreased rate of nonadherence.
More from Dr. Lytle at AAO: How ODs should break bad news to patients
References
1. Brown MT, Bussell J, Dutta S, Davis K, Strong S, Mathew S. Medication Adherence: Truth and Consequences. Am J Med Sci. 2016 Apr;351(4):387-99.
2. DiMatteo, MR. Evidence-based strategies to foster adherence and improve patient outcomes. JAAPA. 2004 Nov;17(11):18-21.
3. Conn VS, Ruppar TM. Medication adherence outcomes of 771 intervention trials: Systematic review and meta-analysis. Prev Med. 2017 Jun;99:269-276.
4. Náfrádi L, Nakamoto K, Schulz PJ. Is patient empowerment the key to promote adherence? A systematic review of the relationship between self-efficacy, health locus of control and medication adherence. PLoS One. 2017 Oct 17;12(10):e0186458.
5. Lin NY, Ramsey RR, Miller JL, McDowell KM, Zhang N, et al. Telehealth delivery of adherence and medication management system improves outcomes in inner-city children with asthma. Pediatr Pulmonol. 2020 Apr;55(4):858-865.
6. Gaffney A, Bor DH, Himmelstein DU, Woolhandler S, McCormick D. The Effect Of Veterans Health Administration Coverage On Cost-Related Medication Nonadherence. Health Aff (Millwood). 2020 Jan;39(1):33-40.
7. Wang MY, Shen MJ, Wan LH, Mo MM, Wu Z, Li LL, Neidlinger SH. Effects of a Comprehensive Reminder System Based on the Health Belief Model for Patients Who Have Had a Stroke on Health Behaviors, Blood Pressure, Disability, and Recurrence From Baseline to 6 Months: A Randomized Controlled Trial. J Cardiovasc Nurs. Mar/Apr 2020;35(2):156-164.
8. Adeghate JO, Hark LA, Brown H, Henderer JD, Waisbourd M, Molineaux J, et al. Philadelphia glaucoma detection and treatment project: ocular outcomes and adherence to follow-up at a single health centre. Can J Ophthalmol. 2019 Dec;54(6):717-722.
9. White DE, Zhao Y, Ogundele A, Fulcher N, Acs A, et al. Real-World Treatment Patterns Of Cyclosporine Ophthalmic Emulsion And Lifitegrast Ophthalmic Solution Among Patients With Dry Eye. Clin Ophthalmol. 2019 Nov 22;13:2285-2292.
10. Miraldi Utz V, Bulas S, Lopper S, Fenchel M, Sa T, Mehta M, et al. Effectiveness of long-term infliximab use and impact of treatment adherence on disease control in refractory, non-infectious pediatric uveitis. Pediatr Rheumatol Online J. 2019 Nov 29;17(1):79.
11. Newman-Casey PA, Salman M, Lee PP, Gatwood JD. Cost-Utility Analysis of Glaucoma Medication Adherence. Ophthalmology. 2020 May;127(5):589-598.
12. McClelland JF, Bodle L, Little JA. Investigation of medication adherence and reasons for poor adherence in patients on long-term glaucoma treatment regimes. Patient Prefer Adherence. 2019 Aug 12;13:431-439.
13. Bilger M, Wong TT, Lee JY, Howard KL, Bundoc FG, Lamoureux EL, Finkelstein EA. Using Adherence-Contingent Rebates on Chronic Disease Treatment Costs to Promote Medication Adherence: Results from a Randomized Controlled Trial. Appl Health Econ Health Policy. 2019 Dec;17(6):841-855.
14. Nishimura K, Tabuchi H, Nakakura S, Nakatani Y, Yorihiro A, Hasegawa S, et al. Evaluation of Automatic Monitoring of Instillation Adherence Using Eye Dropper Bottle Sensor and Deep Learning in Patients With Glaucoma. Transl Vis Sci Technol. 2019 Jun 27;8(3):55.
15. DiMatteo MR, Reiter RC, Gambone JC. Enhancing Medication Adherence Through Communication and Informed Collaborative Choice. Health Comm. 1994;6(4):253-265.
16. Martin LR, Williams SL, Haskard KB, Dimatteo MR. The challenge of patient adherence. Ther Clin Risk Manag. 2005 Sep;1(3):189-199.
17. Eisen SA, Miller DK, Woodward RS, Spitznagel E, Przybeck TR. The effect of prescribed daily dose frequency on patient medication compliance. Arch Intern Med. 1990 Sep;150(9):1881-4.
18. Ashkenazy N, Saboo US, Robertson ZM, Cao J. The effect of patient compliance on remission rates in pediatric noninfectious uveitis. J AAPOS. 2019 Dec;23(6):334.e.
19. Roter DL, Hall JA, Merisca R, Nordstrom B, Cretin D, Svarstad B. Effectiveness of interventions to improve patient compliance: a meta-analysis. Med Care. 1998 Aug;36(8):1138-61.
20. Irvine J, Baker B, Smith J, Jandciu S, Paquette M, Cairns J, et al. Poor adherence to placebo or amiodarone therapy predicts mortality: results from the CAMIAT study. Canadian Amiodarone Myocardial Infarction Arrhythmia Trial. Psychosom Med. 1999 Jul-Aug;61(4):566-75.
21.Shruthi R, Jyothi R, Pundarikaksha HP, Nagesh GN, Tushar TJ. A Study of Medication Compliance in Geriatric Patients with Chronic Illnesses at a Tertiary Care Hospital. J Clin Diagn Res. 2016;10(12):FC40-FC43.
22. National Institutes of Health. Clear communication. Available at: http://www.nih.gov/clearcommunication/. Accessed 1/14/21.
23. Kalichman SC, Ramachandran B, Catz S. Adherence to combination antiretroviral therapies in HIV patients of low health literacy. J Gen Intern Med. 1999;14(5):267-273.
24. Seldon CR, Zorn M, Ratzan S, Parker RM. Current bibliographies in medicine: Health literacy. Available at: https://www.researchgate.net/publication/230877250_National_Library_of_Medicine_Current_Bibliographies_in_Medicine_Health_Literacy. Accessed 1/14/21.
25. Heffer RW, Worchel-Prevatt F, Rae WA, Lopez MA, Young-Saleme T, Orr K, et al. The effects of oral versus written instructions on parents’ recall and satisfaction after pediatric appointments. J Dev Behav Pediatr. 1997 Dec;18(6):377-82.
26. Stewart RB, Caranasos GJ. Medication compliance in the elderly. Med Clin North Am. 1989 Nov;73(6):1551-63.
27. Schillinger D, Piette J, Grumbach K, Wang F, Wilson C, Daher C, Leong-Grotz K, Castro C, Bindman AB. Closing the loop: physician communication with diabetic patients who have low health literacy. Arch Intern Med. 2003 Jan13;163(1):83-90.
28. Rost K, Roter D, Bertakis K, Quill T. Physician-patient familiarity and patient recall of medication changes. The Collaborative Study Group of the SGIM Task Force on the Doctor and Patient. Fam Med. 1990 Nov-Dec;22(6):453-7.
29. Hogbin B, Fallowfield L. Getting it taped: the ‘bad news’ consultation with cancer patients. Br J Hosp Med. 1989 Apr;41(4):330-3.
30. Zolnierek KB, Dimatteo MR. Physician communication and patient adherence to treatment: a meta-analysis. Med Care. 2009 Aug;47(8):826-834.
31. Elwyn G, Edwards A, Britten N. “Doing prescribing”: how might clinicians work differently for better, safer care. Qual Saf Health Care. 2003 Dec;12 Suppl 1(Suppl 1):i33-6.
