Why patients do not adhere to medical advice.

summary

Extent of problem

Taylor (1990) 93% of patients fail to adhere to some aspect of their treatment. Sarafino(1994) People adhere to treatment regimes reasonably closely 78% of the time. Sarafino found the average adherence rates for taking medicine to prevent illness is 60% for short and long term regimes. Compliance to change one's diet or to give up smoking is variable and low.

If you go to the doctors and he or she prescribes tablets that are then not taken, the problem of non-compliance will be exacerbated by the patient lying to the doctor. The doctor will feel that his original prescribed treatment was incorrect and might then prescribe an inappropriate treatment instead.

Compliance with chemotherapy is very high among adults with estimates of better than 90 percent of patients complying with the treatment.

Non compliance takes many forms. Some patients do not keep appointments; others do not follow advice. Many patients fail to collect their prescriptions, discontinue medication early, fail to change their daily routine, and miss follow-up appointments (Sackett and Hayes, 1976).

Kent and Dalgleish (1996) describe a study in which many parents of children who were pre­scribed a ten-day course of penicillin for a strepto­coccal infection did not ensure that their children completed the treatment. The majority of the par­ents understood the diagnosis, were familiar with the medicine and knew how to obtain it. Despite the fact that the medication was free, the doctors were aware of the study and the families knew they would be followed up, by day three of the treatment 41% of the children were still being given the penicillin, and by day six only 29% were being given it. This study very clearly illustrates the problem of non-adherence.

The costs associated with non-adherence can be high. The illness may be prolonged in the patient and he or she may need extra visits to the doctor. These are not the only costs, however, as the person may have a longer recovery period, might need more time off work or even require a stay in hospi­tal. Non-adherence may lead to as much as 10%—20% of patients needing a second prescrip­tion, 5%—10% visiting their doctor for a second time, the same number needing extra days off work, and about 0.25 %—1% needing hospitalization (Ley, 1997).

There are financial implications for non-compliance. In 1980, between $396 and $792 million were wasted in the USA because of non-compliance to prescribed drugs (Ogden, 1996).

Methodological problem of estimating the level of adherence to medical regimes.

Percentages are overestimated because patients who tend to volunteer for these studies would be more likely to be compliant.

Patients often lie about their level of adherence, so as to present a good impression of themselves. It has been reported in the press that those patients who smoke may be afforded a low level of priority, when they are in need of a transplant. Patients might lie about their smoking, to avoid such discrimination.

Why patients do and don't adhere to advice

Patients are less likely to change habits than heed medical advice to take medicine (Haynes, 1976). Patients who view their illness as severe are more likely to comply (Becker & Rosenstock, 1984). Notice it is how the patient views the seriousness of the illness, not what the physician thinks! Doctors tend to blame their patients for non-adherence, attributing their behaviour to characteristics of their patients (mental capacity or personality traits) - Davis (1966). [Can you think how this relates to the Nisbett (Actor-observer effects key study?]. Research has shown that it is not the patient's personality that predicts non-adherence, but a combination of factors arising out of the doctor - patient relationship (e.g. Ley 1982). Factors such as age and gender are predictive of compliance, depending upon what instructions are to be complied with. For example, adolescents are less likely to comply with a long term treatment that makes them appear different from their peers. Classic experiments - Milgram (1963) and Asch (1955. Milgram's experiment demonstrated that ordinary people will obey authority figures, to the extent that they would administer potentially lethal 'electric shocks' to a mild-mannered victim. Asch's experiment demonstrated that people will agree with others even though it is obvious others are wrong.

If medication is prescribed over a long time, it's more likely to be discontinued early (Haynes 1976).

Some patients are complainers and others accept advice more readily. Complainers will always complain and thus be dissatisfied with the doctor, and therefore less compliant. Some patients are predisposed never to follow advice. However there is little evidence for the link between personality and compliance.

Economic factors cannot be excluded; doctors must be sensitive to costs, in terms of both money and time off work (Heib and Wang, 1974).

If several factors are considered, then more accurate predictions of compliance can be forecast. Korsch et al (1978) were able to identify almost 90 percent of non-compliant patients using multiple background variables.

The relationship between patient satisfaction and compliance has been studied by Francis et al. 1969

Patients' Report

Percent Highly Compliant

Doctor businesslike

31

Doctor friendly and not businesslike

46

High satisfaction with consultation

53

Moderate satisfaction with consultation

43

Moderate dissatisfaction with consultation

32

High dissatisfaction with consultation

17

Many patients have serious misconceptions about their illness. For example, a stomach ulcer patient who believes that acid is produced in the stomach each time one eats, is less likely to follow the doctor's advice to eat small, frequent meals.

Types of request

  1. requests for short-term compliance with simple treatments
  2. requests for positive additions to lifestyle
  3. requests to stop certain behaviours
  4. requests for long-term treatment regimes

People's attitudes do not necessarily match their behaviour.

Ley model of patient compliance (1989).

Evaluation

A useful theory or model in health psychology should:

  1. generate significant research
  2. organise and explain observations
  3. help the practitioner predict and change behaviours.

Ley's model has satisfied or of the above criteria, as the following research indicates.

Patient satisfaction

Ley (1988) reviews 21 studies of hospital patients and found that 28% of general practice patients in the UK were dissatisfied with the treatment they received. Dissatisfaction amongst hospital patients was even higher with 41 per cent dissatisfied with their treatment. The dissatisfaction stemmed from affective aspects of the consultation (e.g. lack of emotional support and understanding), behavioural aspects (e.g. prescribing, adequate explanations) and competence (e.g. appropriateness of the referral, diagnosis).

It was found that patients were "information seekers" (i.e. wanted to know as much information is possible about their condition), rather than "information blunters" (i.e. did not want to know the true seriousness of their condition).

Over 85% of cancer patients wanted all information about diagnosis, treatment and prognosis (the chances of treatment being successful) (Reynolds et al., 1981).

60 to 98% of terminally ill patients wanted to know their bad news (Veatch, 1978).

Older research had found that a small but significant group did not want to be given the truth for cancer and heart disease (Kubler-Ross, 1969). These findings could be due, in part, to the attitudes that prevailed during the late Sixties. Research suggests that attitudes have changed since then.

Patient understanding

Boyle (1970) asked patients to define a range of different illnesses and found the following:

Illness to be defined

% correct

Arthritis

85

Bronchitis

80

Jaundice

77

Palpitations

52

 

Patients were also poor at being able to locate various organs.

Organ

% correct

Liver

49

Heart

42

Stomach

20

Roth (1979) found that although patients understood that smoking is causally related to lung cancer, 50% thought that lung cancer caused by smoking had a good prognosis for recovery.

It was also found that 13% of patients thought that hypertension could be cured by treatment when it can only be managed.

Patient recall

Bain (1977) tested recall of a sample of patients who attended a GP practice. The following was found:

Instruction to be recalled

% unable to recall

The name of the drug prescribed

37

The frequency of the dose

23

The duration of the treatment

25

Crichton et al. (1978) found that 22% of patients had forgotten their advised treatment regimes after visiting their GPs.

Ley (1989) found that the following factors increased recall of information:

Age has no effect on recall success.

Compare this list with the one below

Cognitive and emotional factors in patients' recall of information (DiMatteo & DiNicola 1982).

  1. Patients forget much of what is told to them
  2. Instructions and advice are more likely to be forgotten than other information
  3. The more a patient is told the greater the proportion a patient will forget
  4. Patients remember a) what they are told first and b) what they consider to be important
  5. Intelligence is not a factor (but see above)
  6. Age is not a factor
  7. Moderately anxious patients recall more than highly anxious patients
  8. Prior medical knowledge aids recall.

 

Homedes (1991) has reported that more than 200 variables affect compliance. He categorises them as:

The health belief model

Becker and Rosenstock (1984)

  1. Evaluating the threat
    Seriousness and vulnerability are taken into account. Being overweight would make you more vulnerable to a heart attack. A heart attack is serious. The patients relative youth would mean he or she is less vulnerable. And so on. Seriousness and vulnerability being high would be a good predictor of the likelihood of action. However, there are other factors that need to be taken into account. a recent media campaign would be a cue to action. The patient would need to work out the costs and benefits of the treatment as well.
  2. Cost-benefit analysis.
    Will the benefits outweigh the costs? Barriers (or costs) might be financial, difficulty getting to a health clinic, not wanting to admit that they are getting old. Benefits would be improved health, less risk from illness and less anxiety.

Perceptions of severity and susceptibility by the patient are related to compliance (Becker 1976). Patients who believe they are likely to become ill and that this eventuality would have negative consequences are more likely to take some action. Simple beliefs regarding the likelihood that medication will improve the patient's condition are very potent determinants of compliance (Becker 1976). Any question of safety of treatment, side effects, or distress associated with treatment become very powerful suppressers and reduce the likelihood that patients will do as they were told (Becker 1974).

Actual severity of an illness is not related to compliance, but patient perception of severity is.

Abraham et al (1992) studied 300 sexually active Scottish teenagers. The seriousness of AIDS and the perceived vulnerability of contracting the illness were not the factors that influenced the teenagers. The awkwardness of use and the likely response from their partner, were seen as costs that outweighed the benefits. The teenagers therefore tended not to use condoms! It would make sense to concentrate advertising campaigns on the barriers to condom use.

It is difficult to assess the health belief model as it is difficult to measure variables such as perceived susceptibility. Habits, such as cleaning your teeth are not easily explained by the model. The model has limited predictive value, but can be useful when trying to explain somebody's behaviour.

The Health Belief model is a comprehensive model. Revisions in the model have expanded its range to include intentions as well as beliefs (Becker 1974).Other models that are less comprehensive are the theory of reasoned action, protection motivation theory, Naive health theories and subjective expected utility theory.

Naive health theories.

Patients often develop their own incorrect theories about their illnesses. Such theories develop because a particular behaviour has become erroneously associated with an improvement in their condition. Such beliefs interfere with the understanding of the doctor's instructions. The instructions are interpreted so as to accord with their naive health theory (Bishop and Converse, 1986).

The model has two strengths. One is that it explains why a patient who intends to comply actually does not. Secondly, the model is easily testable.

Rational non-adherence

Sometimes the side effects of a treatment can be so devastating, that the patient decides, quite rationally, not to proceed with the treatment. Bulpitt (1988) medication used for the treatment of hypertension reduced the symptoms of depression and headache. However, the men taking the drug experienced increased sexual problems (difficulty with ejaculation and impotence). Chapin (1980) suggested that 10% of admissions to a geriatric unit were the result of drug side effects. Most non-adherence in arthritis patients was owing to unintentional reasons (e.g. forgetting); the common intentional reasons were side effects and cost (Lorish et al, 1989).

Reasons for non-adherence (Sarafino 1994)

  1. believe treatment is not working
  2. Unpleasant side-effects
  3. Confusion over when to take the take the medicine and how much to take.
  4. Cost of the treatment, or other practical barriers
  5. To check whether illness is still present, when medication is stopped.

Other useful concepts

  1. Behavioural explanations - habits, imitation (young smokers copying peers), reinforcement (short term treatment will provide this, but long term treatment would not).
  2. Defence mechanisms - e.g. smokers might use avoidance by avoiding information about the harmful effects of smoking. Also, they could use denial, pretending that smoking is harmless.
  3. Conformity - e.g. men acting hard in front of their mates, and therefore not complying with their doctor's requests.
  4. Self-efficacy (believe they can do something about the problem) and locus of control (feel that they have some control over the illness).

Self-efficacy and conformity

It is better to have a self-efficacy that is slightly higher than one's true ability. This will give one the confidence to undertake a task and make them resilient to giving up. We assess our self-efficacy by:

  1. Observing others
  2. Social and self-persuasion
  3. Monitoring our emotional states

Payne and Walker (1996) suggest that because people who have low self-esteem and low self-efficacy do not value their own ideas, they are more likely to adhere to medical advice because they value what the doctor tells them more than someone with high self-esteem. Similarly, a number of factors have been shown to be important in predicting the likelihood that a person will conform (or adhere), and one of these is if a person perceives that someone else has greater expertise than them (i.e. has a ‘powerful others’ locus of control). This is often true of how people regard doctors.

Another factor that affects conformity is that people are more likely to conform if they have relatively little information on which to base their judgment the less information, the higher the rate of conform­ity. This may seem counter-intuitive, but combined with some of the other factors already mentioned. Payne and Walker (1996), argue that many situations relating to patient adherence in health care are like this. Doctors tend to explain things to their patients on a ‘need to know’ basis, which leaves patients with little real knowledge about their condition or treatment.

 

Poradzisz (2002) examined the influence of selected variables on quality of life (QOL) and adherence to diabetes regimen in a convenience sample of adults with Type 2 diabetes. Participants were all clients of the same American Diabetes Association-recognized diabetes centre and had previously received comprehensive diabetes education. Sixty-one percent of the participants reported education beyond the high school level; 41.5% reported annual family income of more than $50,000. In Part 1 of the study, participants (n = 94) completed a survey that included measures of self-efficacy, illness demand, family support, satisfaction with care, QOL, and adherence. In Part 2, selected respondents (n = 23) participated in an interview that focused on the same variables. For both survey and interview data, comparisons were made between those in the high and low adherence groups with regard to self-efficacy, illness demand, family support, and satisfaction with care. Persons in the low adherence group, as well as those with longer duration of diabetes, had significantly lower self efficacy and higher illness demand than those in the high adherence group. Individuals in both high and low adherence groups reported lower levels of self-efficacy when it was necessary to make adjustments in the diabetes self-care regimen. Self-efficacy was found to be a significant predictor of both adherence (R2 = 0.35, p <.001) and QOL (R2 = 0.40, p <.001). Illness demand and family support contributed to QOL, but did not significantly influence adherence for this sample. A moderate correlation (r = 0.50, p <.001) was found between adherence and QOL. Data from the interview participants indicated that, while strict adherence was generally perceived as having a positive effect on QOL, low adherers had a tendency to perceive more negative effects. The results suggest that diabetes educators should focus on efforts to improve self efficacy and the individual's perception of QOL in order to enhance adherence to diabetes regimen, and that ongoing attention to these aspects is needed in order to provide optimal support to those with longer disease duration.

Diabetes and Compliance

Pancreas produces the hormone insulin, which is responsible for the storage and use of glucose. Diabetics either produce no insulin (Type I), or do not produce enough insulin (Type II). Type I would require insulin replacement by injection. Type II requires a good diet, weight management and oral medication. Too little glucose in the blood would lead to a hypoglycemic attack ('a hypo'), which is life threatening. Too much glucose produces little immediate effect (unless the patient has too much glucose over a long period of time). The long term effects of too much glucose is likely to be circulation and heart problems (Bradley 1994).

Diabetics have to respond to the following health requests:

  1. Give themselves injections of insulin, on a regular basis, using the correct dosage and in a hygienic way.
  2. Regularly test for the level of glucose in the blood.
  3. Eat on a regular basis and regulate the intake of carbohydrates
  4. Visit a health centre for a general check up on a regular basis.
  5. Take regular exercise.
  6. Avoid alcohol, as this lowers the level of blood sugar.

Wing et al 1986 Table of non-compliance of diabetics

The non-compliance of diabetics

Unhygienic injections

80%

Wrong insulin dose

58%

Inappropriate diet

75%

Irregular diet

75%

Glasgow et al (1987) - Diabetics have more difficulty following exercise and diet advice. Data was collected by the patients own 'self-report'. This method is unreliable. Glucose-testing machines fitted with a memory chip showed that patients were inaccurate in their self-reporting.

Possible reasons as to why diabetics do not comply might be:

  1. Embarrassment of testing and injecting in social settings.
  2. Lack of social support.
  3. Belief that they do not need to test for glucose level, as they can tell anyhow.

 

Swigonski (1987) found that social support in kidney disease patients can have adverse effect. Such patients were more likely not to comply with limiting their fluid intake. This is because social occasions often involve drinking!

Bradley (1994) found evidence to support the view that patients can be aware of their glucose levels:

  1. Some diabetics were very accurate at being able to do this.
  2. Some believe they have this ability, but are not without training.
  3. Symptoms tend to be reliable for the individual, but vary between individuals.
  4. Most diabetics can learn to be aware of their blood-glucose levels. (A good example of the application of psychology).

Kutz  (2000) examined the role that a number of psychosocial variables may have played in affecting adherence behaviours in two populations of patients with Type 2 diabetes mellitus. The studies' psychosocial variables included the patient-practitioner relationship (PPR), coping style, social support, psychological distress, perceived stress, and SES. Ultimately, it was discovered that when individuals were satisfied with interpersonal aspects of their medical care they were receiving in their endocrinology clinics, they were more likely to adhere to treatment recommendations. Furthermore, it was discovered that when patients were less likely to utilize avoidance as a means by which to cope with their illness, they tended to adhere more to prescribed treatment regimens. It was also indicated that general satisfaction with medical care, as well as the perceived level of social support the patients felt they obtained, contributed to adherence practices.

Shillitoe and Miles (1989) in defence of the diabetics against unjust criticism:

  1. They do not have a 'specific' set of guidelines to follow.
  2. They follow some instructions and not others.
  3. Treatment varies from day to day, depending on many different factors (e.g. exercise, infection, etc.). This makes it difficult to define and measure compliance.

'Compliance' seems to be the wrong word. Perhaps 'adherence' or 'levels of self-care behaviour' might be more appropriate, depending on the situation.

Old People and compliance

Kaplan et al (1993)

  1. Difficulty in understanding and following complex instructions.
  2. Difficulty in opening medicine bottles (e.g. child-proof caps).
  3. Range of medicines increase the risks of side-effects.

According to a new study published in the journal Stroke, high blood pressure actually shrinks brain tissue and speeds memory loss and cognitive dysfunction. Brain atrophy and a reduction in memory were noted in hypertensive patients as compared to non-hypertensive patients. The patients involved, had not had a stroke and had no other diagnosed medical conditions.