Abstract
Asthma is a potentially lethal, frequently disabling, and almost invariably distressing disease. Adequate management relies heavily on the patient's ability to detect changes in breathing, assess them accurately, and respond to them appropriately. Disturbances in emotional state are very likely to have impact on this process. As attention to education for health workers has been successfully undertaken, and improved medications and techniques for managing asthma have been mastered, the focus for decreasing mortality and morbidity in asthma has increasingly been on patient knowledge and ability to put this knowledge into practice. Patient education, the introduction of peak flow meters and daily measurement of peak flows have all been initiated enthusiastically. Results have been somewhat disappointing. Apparently even well-educated patients who possess written action plans cope poorly with their disease [1–3].
Increasingly, psychological, psychosocial and attitudinal factors have been studied as possible contributors to illness and death in a variety of diseases [4–7], as well as asthma [8]. Hypothesised pathways include effects on lung function directly [9], on symptom reporting [10],[11], on healthcare use [12] and compliance with treatment [13]. In addition, socioeconomic and psychological factors have been shown to impact on the course of asthma and in particular to act as risk factors for poor outcome [12],[14],[15].
Anxiety, fear and panic have consistently attracted the greatest interest. The panic-fear symptom constellation has been shown to occur at higher rates in people with asthma compared to non-asthmatic subjects [16],[17]. This pattern has been measured as a response to the experience of breathlessness in one study [18]. It may also reflect a non-specific personality response style evoked by a range of conflicts and difficulties in addition to breathlessness [19]. In addition, anxiety, and in particular its tendency to be associated with abnormal breathing patterns and sensations of air hunger secondary to hyperventilation, may itself be a source of misidentification of bodily states, leading to inappropriate decision-making. Finally, both fear and anger are accompanied by autonomic responses that might be expected to increase alpha sympathetic activity and vagal blocking, thus exacerbating asthma at a psychophysiological level [20–22].
This study describes and examines some parameters of psychosocial wellbeing in a group of asthmatic patients who are at high risk of serious morbidity (i.e. those in whom potentially harmful influences are the most worrisome), examines their interrelationships and examines the hypothesis that these factors will be predictive of asthma self-management knowledge.
Method
Study population
Volunteers aged 14–76 diagnosed as having asthma with clearly demonstrable reversible airflow obstruction attending a hospital-based asthma clinic for their regular appointment were approached to take part. The clinic caters for patients with moderate to severe asthma and the majority of referrals are as a result of recurrent hospital admissions or emergency room (ER) attendances. Since 1982, all patients presenting with severe life-threatening asthma and admitted to an intensive care unit (ICU) have been automatically referred to the asthma clinic. As well as respiratory physicians, the clinic is staffed by a full-time asthma nurse educator who reviews patients at each visit and is available for follow-up in the community, and a half-time clinical psychologist. Thus, patients under the clinic's care could be considered to be receiving a comprehensive multidisciplinary package of care for their condition.
Over the 3 months of the study, there were 386 appointments in the clinic. Of the 298 who attended, 235 were approached to participate. Eight (3.4%) did not in fact have asthma, 113 (48%) said they had no time available, 13 (5.5%) had language difficulties, 18 (7.7%) were too sick, and 13 (5.5%) refused to take part. The final sample of 80 subjects thus represented 34% of those approached. The mean age was 41.2 years (SD=15.4); 43.8% were male, 88.8% were European, 5.0% were Maori, 3.8% were Pacific Islanders and 2.5% were other. Of the group who did not take part, there was no significant difference in terms of age or gender. While there was no overall difference in race, the subgroup who were found to be unsuitable because of language difficulties was predominantly Pacific Islander.
A trained interviewer was used both to approach and explain the study, and then to conduct the interview, which comprised a combination of questionnaires and structured interviews in the Social Support and Asthma Knowledge Instruments.
Instruments
Anxiety and depression
The Hospital Anxiety and Depression Scale (HAD) [23], which is specifically tailored towards eliciting distress in physically ill subjects, is well validated against other depression scales, but is not asthma-specific. Responses are scored from 0 to 3 with patients selecting the description which best suits them for each item. A cut-off of ≥ 8 was taken to indicate possible caseness, and > 10 definite caseness.
Practical asthma knowledge
This was assessed by the subject's description of the action they would take in response to each of two hypothetical evolving attacks. This scenario method is described in detail elsewhere [24].
Social support
This scale [25] was originally designed for use in patients with cardiovascular disorders. Angina and myocardial infarction present similar challenges to asthma in its ongoing and life-threatening manifestations and, therefore, modification for use in this study was undertaken. The scale measures both general support, and support specific for illness, such as the giving of information or advice, offering help in management of the condition on a day-to-day basis, availability and adequacy of support during an acute attack, and level of conflict with the person identified as the supporter. All responses are scored from 1 (= None) to 5 (= A great deal).
Adjustment to asthma
This was measured using a modification of the ‘Attitudes to Asthma Questionnaire’ [26],[27]. This 33-item questionnaire, developed in a sample of English outpatients, includes items relating to feelings about asthma, perceived effects of asthma on self and family, attitudes to the condition and to medication used in its treatment and the doctor's supervising management. Responses are scored on a Leikert-type scale; however, there is no ‘neutral’ response allowed. The questionnaire was modified by the deletion of several ambiguous questions, by re-wording of several questions, by removal of headings used in the original layout grouping items, and by re-arrangements of items. In addition, a psychological interview was undertaken in a subsample to test consensual validity.
Ethics
All subjects gave written informed consent to participate in the study which was approved by the Auckland Area Health Board Ethics Committee.
Statistical analysis
Data are expressed as mean (± SD). A p-value of < 0.05 was regarded as significant. Data were analysed using multivariate regression analysis; factor analysis was used for the Attitudes to Asthma Questionnaire.
Methods
Morbidity
Subjects had attended the clinic on an average of 9.8 occasions in total; 3.8 in the previous year. Twenty-nine percent had been admitted to hospital in the previous year for acute asthma and over 60% had experienced a life-threatening attack as defined by Richards et al. [28]. They had been on 2.44 (± 2.35) courses of oral corticosteroids in the past year, and had spent a mean of 54.07 (± 90.46) days on oral steroids in the same year. Current medications included inhaled corticosteroids in all patients as well as beta-agonists. They were thus a group with at least moderately severe asthma.
Mood disturbance
Emotional distress in this sample varied according to the specific symptom sought. Twenty-five percent had possible and 19% had definite caseness for clinically significant anxiety as measured by the HAD scale (mean score = 7.4 ± 4.2). In contrast, only 10.3% had possible and 1.5% definite depression (3.4 ± 2.9).
Social support
Social support for general decision-making and personal worries was available for 73%. Most commonly, the person named was a spouse or a parent. Friends were frequently named as showing concern, but not as being consulted for worries or decision-making. Eighty-five percent of the sample named someone who helped them with their asthma; 23% named more than one person. Nominated persons in this group were mostly spouse or parent, with friends and neighbours featuring less commonly. The other 15% either had no-one available or preferred to look after their asthma themselves. Subjects were less likely to receive information and advice on a day-to day basis than to have help during an attack. During an acute attack, 96% had help available, 97% were satisfied with this support and 91% reported no major conflict between themselves and the support person in the previous year. The data were categorised by division into a group rated as having inadequate support for any of the following reasons: those who could name no support person, those whose person was unavailable when an acute attack occurred; those whose help was unsatisfactory; those whose support person had created high levels of conflict in the previous year. The remaining group was classed as having adequate support. Using this definition, 24% had inadequate support.
Attitudes to asthma
The 33 items of the Attitudes to Asthma Questionnaire was subjected to factor analysis. This resulted in a three-factor solution similar to the one described by others [29], which accounted for 84.8% of the variance. Factor 1, named Emotional Maladjustment to Asthma loaded with 11 items such as ‘I worry that I might die from asthma’, ‘Asthma has made me less physically attractive’, ‘Even when I feel well I worry about getting an attack of asthma’. This factor was significantly predicted by anxiety and depression, by age and weakly by race, as well as by number of courses of steroids (but not number of days on steroid) in the previous year. Factor 2, labelled Doctor-Patient Relationship, contained items like ‘I feel understood by my doctor’, ‘I have confidence in my doctor's management of my asthma’. This factor was weakly predicted by depression only. Factor 3, labelled Asthma-Related Stigma loaded with items such as ‘I feel somehow to blame for being an asthmatic’ and ‘It embarrasses me to use my inhalers in public’. Again, this factor was predicted only by depression score.
Asthma management knowledge
The measurement of practical knowledge about asthma showed poor performance on both the scenarios depicting a slow onset attack, and a rapid onset attack. Mean score out of a maximum possible of 25 was 12.76 (± 3.95) and 13.93 (± 4.78), respectively. A score of 15 or more was considered to represent satisfactory practical self-management knowledge [24]. Further details on performance on items within the scenario will be reported elsewhere. The scores for the rapid onset and slow onset hypothetical attacks were each predicted by the other, and the score for the slow onset attack was predicted by whether such an attack had ever been experienced by the patient; this did not hold for the rapid onset attack. There was no relationship to sex, age, race, or morbidity data.
The hypothesis that practical asthma knowledge would be related to psychosocial factors was not confirmed: neither score was predicted by any of the Attitude Factors, nor all three combined in a model, nor by anxiety, depression, or social support.
Discussion
This was a pilot study, and contains many shortcomings. The low response rate, while not surprising in the setting of an outpatient clinic population with other priorities, was a possible source of data distortion. Apart from race, which in New Zealand has some predictive value for socioeconomic status, no data were gathered on sociodemographic variables. It is likely that educational and occupational status as well as home living conditions all have significance for asthma management ability. Our sample, comprising patients who managed the often difficult task of keeping a hospital appointment, is likely to represent Europeans of higher socioeconomic class who are more compliant and reliable [30].
Given the intensive level of education provided in the Asthma Clinic, it was surprising, disappointing, and of great concern that the level of practical knowledge of asthma self-management was so low. The mean score on both kinds of ‘attack’ fell below the level considered acceptable, suggesting that most of these patients still have poor grasp of what to do in the event of worsening asthma. Items poorly handled included an apparent inability to seek medical help when asthma was deteriorating, or to call for an ambulance when hypothetically unable to speak or move because of breathlessness. It is precisely such failures to act which have been implicated in asthma deaths. While the correlation between knowledge assessed in this way and actual behaviour is uncertain, studies have shown that scores are likely to be strongly predictive [31].
None of the dimensions measured in this study influenced the level of practical knowledge of asthma self-management, refuting the original hypothesis. If psychosocial factors do adversely influence asthma morbidity, then it appears that their effect is not on the acquisition and retention of asthma knowledge. Whether they have an impact on actual behaviour during an attack remains unknown.
The level of psychiatric distress was high, suggesting a need for attention in its own right, however. In line with other work, the most common form of psychological morbidity was anxiety. Comparing this study to another by using the same instrument [32], the most striking difference was in the rate of definite cases of clinically significant anxiety. This was found in three times as many of our subjects (19% vs 6%) as in the other study. By comparison, the rates of milder anxiety symptoms of possible cases were about the same in the two populations (25% vs 24%).
However, the sample in the Janson et al. study [32] was drawn from a community-based population, with the most severely affected patients being defined as those with ‘nocturnal breathlessness, attacks of asthma, or current use of medication’, much lower morbidity than in our sample. This suggests that while states of fear of a mild sort are common in all those who suffer from asthma, whether at a mild or more serious level, serious anxiety may be a problem in the population with more clinically serious airflow obstruction.
This study does not, of course, clarify whether it is the severe states of anxiety which makes the asthma hard to control, the extremely frightening consequences of living with a life-threatening disease which creates anxiety, or a mixture of the two. Nor does it examine the possible role of medication in creating or maintaining states of anxiety.
Our results, however, underline the need to address anxiety in asthma management. As has been alluded to, excessive anxiety can itself induce dysfunctional breathing patterns including the sensation of air hunger associated with hyperventilation which may then be mistakenly attributed to airway obstruction [33]. Additionally, high levels of anxiety may impair the ability to make correct self-management decisions. Finally, carers may be influenced by excessive anxiety and may alter medical regimes inappropriately [34]. However, asthma is an illness which demands constant vigilance for the early warning signs of a serious attack. Suppression or denial of anxiety, while protective in terms of unpleasant emotion, may be in itself life-threatening. Yellowlees et al. [35] described a group of asthmatics with almost no anxiety despite having had experience of a life-threatening attack. Clearly, such attitudes may impair self-management, and make a true therapeutic alliance between health professional and patient difficult [36]. Because our study described patients between episodes, when there was no need for ‘signal’ anxiety, it is not possible to examine whether these patients may have inappropriately low levels of anxiety at times of crisis.
By contrast, we did not find depression to be a major problem in this group. The rate of definite caseness for depression (1.3%) was almost exactly the same as the Janson et al. study [32]; however, the rate of possible caseness (10%) was higher than their finding of 7%. Both of these studies had rates which were lower again than those of Garden and Ayres [37], who found 15% of their sample of so-called ‘brittle’ asthmatics to have current mood disorder; this increased rate of depression may relate to the apparent inability of health professionals to improve the course of illness in this group of patients. Prevalence of depression in medical patients has been variously documented between 5% and 20%, depending on the population chosen and the instrument used [38]. Generally, the rate has been shown to be higher in the more seriously ill [39], and higher in illnesses affecting the central nervous system [40]. Drugs also affect the rate. Given the association between corticosteroids and depressive symptoms [41], and the high level of morbidity of our patients, our overall rate of probable or definite caseness of just over 11% was consistent with these findings. Earlier studies have suggested a highly significant relationship between depression and severe asthma [42] and asthma death [15],[43], and have implied much higher rates of depression. This may relate either to the imprecision with which psychiatric symptoms are measured and described or to the different patient populations studied.
The factor structure of the Attitudes to Asthma Questionnaire, with a three-factor solution, provides some interesting insights. The separate factor accounting for items dealing with doctor-patient relationships suggests that high emotional distress relating to asthma does not per se interfere with the development of a good therapeutic relationship. Conversely of course, such a relationship is not in itself sufficient to alleviate distress, suggesting the need for psychological symptoms to be specifically targeted. Factor 1, containing items relating to worry and distress, remains independent of items implying stigma and shame which loaded on Factor 3. It remains to be seen how these clusters of emotional factors relate to the outcome of an asthma attack.
It is noteworthy that the factor structure obtained in this group of patients was very similar to that described by Sibbald in a sample of outpatient, and apparently less severe, asthmatics [29]. In her study, there was a significant correlation between each of the factors and morbidity. Our findings did not replicate this, perhaps because our measures of morbidity (i.e. use of hospital facilities and use of oral steroids) were deliberately chosen to be as independent as possible of patient perception of illness. This was intended to minimise the potential confounding effect of psychological state on symptom perception and to use instead the view of health care professionals into level of morbidity. Interestingly, there is evidence that doctors are themselves affected by patients' anxiety concerning their symptoms [34].
Further work is under way to investigate psychosocial factors in the context of a developing attack of asthma to help sort out the temporal sequence in an attack, and in particular to elucidate at which points psychosocial factors play their part.
Footnotes
Acknowledgements
This research was supported by grants from the Health Research Council and Lottery Health Research of New Zealand.