IS STRUCTURED ALLERGY HISTORY SUFFICIENT WHEN ASSESSING PATIENTS WITH ASTHMA AND RHINITIS IN GENERAL PRACTICE? Helen E. Smith, DM, FFPHM, Claire Hogger, Camille Lallemant, MSc, David Crook, PhD, and Anthony J. Frew, MD, FRCP A study in which one hundred twenty-seven patients (ages 18-55) were identified for diagnoses of rhinitis, asthma, or both; the objective being to examine the extent to which formal allergy assessment (structured allergy history, skin prick test, and written patient education/advice tailored to allergic response), improves accuracy of diagnosis when compared to patient selfreport or structured allergy history alone. Patient s allergy status was determined by patient self opinion and on structured allergy history (a standard questionnaire including demographics, diagnoses, symptoms, duration of disease(s), age of onset, exacerbating factors, seasonal variations, effect on lifestyle, home allergens, family history of atopy, and current management, i.e. medications, allergen avoidance and desensitization) alone and then compared to a formal allergy assessment. Both included skin prick tests to 5 common aeroallergens (grass pollen mix, tree pollen mix, house dust mite, cat fur, and dog hair). Self- reporting led to false allergic status in many patients. Structured allergy history by itself was somewhat better but also resulted in false positive results for cat allergy, grass pollen, house dust mite, tree pollen, and dog (32%, 48%, 75%, 54% and 27%, respectively) when compared with formal allergy assessment. Findings were that the combined structured allergy history in conjunction with skin prick test provided a much superior and accurate assessment than the patients own assessment of their allergic status or a structured history alone.
Is structured allergy history sufficient when assessing patients with asthma and rhinitis in general practice? Helen E. Smith, DM, FFPHM, a Claire Hogger, b Camille Lallemant, MSc, a David Crook, PhD, a and Anthony J. Frew, MD, FRCP c Brighton and Southampton, United Kingdom Background: Many United Kingdom patients with asthma and rhinitis are allergic, but in primary care few diagnostic and management decisions are made with formal allergy assessment. Arguably, knowing a patient s atopic status might be helpful in distinguishing the cause of disease and in selecting appropriate treatments. Objectives: Our objective was to estimate the extent to which a formal allergy assessment (a structured allergy history and skin prick tests to 5 common aeroallergens) would improve the precision of allergy diagnosis compared with a patient s selfreport or the structured allergy history alone. Methods: One hundred twenty-seven patients with asthma, rhinitis, or both were recruited from 4 general practices in Wessex, United Kingdom. Allergy status based on the patient s opinion and on structured allergy history alone was compared with formal allergy assessment. Assessments were validated by an independent allergy specialist reviewing the files. Patients were given written advice specific to their allergies and followed up 3 months later to assess satisfaction, recall, and effect on health and behavior. Results: Self-reporting misclassified allergic status in many patients. A structured allergy history alone was little better and resulted in false-positive rates for cat allergy of 32%, grass pollen of 48%, house dust mite of 75%, tree pollen of 54%, and dog of 27% compared with formal allergy assessment. Skin prick testing combined with a structured history was essential to reach a correct causative diagnosis. Three months later, 41% patients had made changes to lifestyle, medications, or both, and 18% reported clinical improvement. Conclusions: Skin prick testing improves the accuracy of an assessment of allergic status based on patient opinion or a structured allergy history alone. (J Allergy Clin Immunol 2009;123:646-50.) From a the Division of Primary Care and Public Health, Brighton and Sussex Medical School, Brighton; b the School of Medicine, University of Southampton; and c the Department of Respiratory Medicine, Brighton General Hospital. ALK-Abelló provided the allergens and test materials free of charge. Disclosure of potential conflict of interest: A. J. Frew has received honoraria from and served on advisory boards for ALK-Abelló, has received speakers honoraria and research support from Allergy Therapeutics, and has served as an officer for the European Academy of Allergy, Asthma & Immunology and the British Society for Allergy and Clinical Immunology. The rest of the authors have declared that they have no conflict of interest. Received for publication April 15, 2008; revised October 21, 2008; accepted for publication November 7, 2008. Available online January 9, 2009. Reprint requests: Helen Smith, DM, FFPHM, Division of Primary Care and Public Health, Brighton and Sussex Medical School, Mayfield House, Falmer, Brighton BN1 9PH, United Kingdom. E-mail: h.e.smith@bsms.ac.uk. 0091-6749/$36.00 Ó 2009 American Academy of Allergy, Asthma & Immunology doi:10.1016/j.jaci.2008.11.005 Key words: Allergy history, primary care, skin prick tests, validity Atopic diseases are a common reason for consulting United Kingdom (UK) general practitioners (GPs), representing up to 6% of all consultations and accounting for 10% of primary care prescribing costs. 1 Many patients with asthma or rhinitis symptoms have allergies, 2 but in primary care diagnostic and management decisions are usually made without taking a formal allergy history or performing skin prick tests (SPTs). 3 Failing to distinguish between allergic and nonallergic causes leads to empiric decisions on treatment. Advice regarding allergen avoidance is either not given or, if given, is nonspecific to the patient s problem and often therefore futile. Integration of SPTs into general practice has been predicted to improve symptom control and also to increase the cost-effectiveness of managing atopic disease through appropriate targeting of medication and allergy avoidance advice. 4,5 SPTs to common aeroallergens are safe 6 and feasible in a community setting but are rarely offered in UK general practice. One reason is cost: GPs have to pay for the test reagents themselves. 7 Another factor might be the disappointing results of one published study of nurse-led SPTs, which failed to achieve the expected improvement in diagnostic precision because the nurses seemed to make their decisions simply on the basis of the patients skin reactivity rather than integrating these data with the clinical history. 5 In that study the number of planned allergen avoidance interventions decreased by only 9% after SPTs. However, further analysis revealed that if avoidance interventions had been targeted entirely appropriately (ie, to those patients with both a positive SPT response and a positive clinical history), then the number of avoidance measures could have been reduced by more than 50%. If UK GPs are to incorporate SPTs into their daily practice, they will need convincing that a structured allergy history and SPTs can improve on the patient s or GP s unsupported assessment of his or her allergic status and that such interventions are acceptable to the patient. METHODS We identified adult patients (18-55 years of age) given diagnoses of rhinitis, asthma, or both from the computerized records held by 4 general practices in Wessex, UK. To ensure that their problems were still active, we excluded anyone who had not requested a problem-related consultation within the preceding year. Those with a history of anaphylaxis (but not mild angioedema or urticaria) were excluded, as were those who had had an SPT within the preceding 2 years. The patients were invited by letter from their practice to participate in the study. They also received a patient information leaflet and a reply slip with a stamped addressed envelope. Those agreeing to take part responded to a short questionnaire given by telephone to confirm their eligibility and were then given an appointment for allergy assessment. Written consent to participate in the study was obtained at the time of this assessment. Before commencing the allergy assessment, patients were asked to what they 646
J ALLERGY CLIN IMMUNOL VOLUME 123, NUMBER 3 SMITH ET AL 647 Abbreviations used GP: General practitioner SPT: Skin prick test UK: United Kingdom thought they were allergic. The formal allergy assessment consisted of a structured allergy history and SPTs, followed by provision of written advice specific to the patient s allergic status. There were 319 eligible patients, and with an anticipated acceptance rate of 85%, this resulted in a projected sample size of 256 patients. This sample would have allowed estimates of sensitivity and specificity with 95% CIs of 68%. We recruited 127 patients, which resulted in 95% CIs of up to 614%. Structured allergy history A standard questionnaire was administered by the practice nurse to obtain demographic details (age, sex, and employment), clinical details (diagnoses, predominant symptoms, duration of disease or diseases, age of onset, seasonality of symptoms, and exacerbating factors), effect on lifestyle, allergens in the home, family history of allergic disease, and current management (pharmacologic, allergen avoidance, and desensitization). TABLE I. Demographics, pet ownership and disease status of the 127 participants Age (y), mean 6 SD 38.2 6 8.9 Sex Male 70 (55.1%) Female 57 (44.9%) Pet ownership None 53 (41.7%) Cat only 21 (16.5%) Dog only 16 (12.6%) Other pet only 14 (11.0%) Cat and dog 5 (3.9%) Cat and other pet 5 (3.9%) Dog and other pet 8 (6.3%) Cat, dog, and other pet 5 (3.9%) Current diagnosis Rhinitis only 33 (26.0%) Asthma only 15 (11.8%) Rhinitis and asthma 79 (62.2%) SPTs We tested for reactivity to grass pollen mix, tree pollen mix (early and midseason), Dermatophagoides pteronyssinus (house dust mite), cat fur, and dog hair using standardized allergen extracts (ALK-Abelló, Hungerford, UK). These allergens were chosen because they are the ones most frequently responsible for allergy in the UK. Positive (histamine) and negative (saline) controls were also used. Practice nurses were trained in use of SPTs by an allergist (AJF), and SPTs were performed in a standardized way on the volar surface of the forearm by using 1-mm prick lancets (ALK-Abelló) to improve reproducibility. 8 Wheal diameters were recorded after 15 minutes as the arithmetic mean of the maximum diameter and perpendicular diameter. A mean wheal diameter of greater than or equal to 3 mm was considered a positive result (99% specificity). 9 A record of the wheals was obtained by outlining with a fine black roller-tip pen and using clear adhesive tape to transfer the outline directly to the data collection form to enable reliability testing. Advice on allergen avoidance techniques Participants were advised of their allergies and informed of appropriate methods of allergen avoidance by means of allergy-specific leaflets produced and used by the local secondary care allergy clinic. Assessment of validity of decisions about allergic status For all participants, the structured allergy history forms and the recorded SPT responses were reviewed by a specialist in allergy (AJF) who was blinded to the interpretation of the nonspecialist. The decisions about the allergic status of the patient made by the nonspecialist and specialist were compared by using percentage level of agreement (overall, positive, and negative) and the Cohen k statistic. 10 Follow-up A brief questionnaire was posted to the study participants 3 months after their SPTs to assess patient acceptability of the technique and the effect of the allergy assessment on disease management (medication and lifestyle) and symptom control. A stamped addressed envelope was provided for reply. Analysis Data were coded and entered into the SPSS system (SPSS, Inc, Chicago, Ill) for analysis. Descriptive statistics were used to characterize the study FIG 1. Prevalence of allergies based on patients self-reports compared with formal allergy assessment (structured history plus SPT). population. The prevalences of allergies identified by the patients themselves, by means of history alone, and by means of allergy assessment (history and SPTs combined) were compared with each other. The characteristics of the diagnostic approaches (sensitivity, specificity, positive predictive value, negative predictive value, and false-positive rate) were calculated by using the Microsoft Excel spreadsheet DAG-Stat (Microsoft, Redmond, Wash) 11 for allergy ascertainment based on the patient s self-report and of structured history alone by using allergy compared with the gold standard of allergy assessment.
648 SMITH ET AL J ALLERGY CLIN IMMUNOL MARCH 2009 TABLE II. Prevalence of allergies and predictive validity of the 3 different diagnostic approaches (self-report, structured history only, and formal allergy assessment) Allergen Grass pollen Tree pollen Cat Dog House dust mite No. (%) of patients self-reporting allergy 60 (48) 46 (37) 58 (46) 34 (27) 56 (45) Sensitivity (95% CI)* 66 (54-77) 57 (42-71) 81 (67-91) 68 (43-87) 58 (46-70) Specificity (95% CI)* 74 (60-84) 77 (66-86) 75 (64-84) 80 (72-87) 71 (57-82) Predictive value of positive/negative test response (%)* 75/65 63/73 66/87 38/93 70/59 False-positive/misclassification rate (%)* 26/30 23/31 25/23 20/21 29/36 No. (%) of patients with allergy based on structured history (positive 96 (76) 89 (70) 58 (46) 46 (37) 109 (87) allergy history) Sensitivity (95% CI)* 99 (92-100) 94 (84-99) 98 (89-100) 89 (67-99) 97 (90-100) Specificity (95% CI)* 52 (38-65) 46 (35-58) 68 (57-78) 73 (63-81) 25 (15-38) Predictive value of positive/negative test response (%)* 71/97 54/92 65/98 37/98 60/88 False-positive/misclassification rate (%)* 48/23 54/35 32/21 27/25 75/37 No. (%) of patients with allergy based on formal allergy assessment (positive allergy history and positive SPT response) 68 (54) 51 (40) 47 (37) 19 (15) 67 (53) *Compared with formal allergy assessment. FIG 2. Effect of adding SPTs to the structured history on the prevalence of allergies compared with relying on the structured history alone. RESULTS Of the 319 patients invited to participate in this study, 137 consented, but 10 were ineligible or failed to attend for assessment, resulting in a total of 127 participants. The majority of participants (62%) reported both asthma and rhinitis (Table I). Fifty-eight percent lived in households with pets. Before assessment, 101 (80%) patients thought they were allergic to 1 or more of the aeroallergens being investigated. When they were assessed with the structured allergy questionnaire, 124 (98%) provided a history suggestive of allergy to at least 1 aeroallergen. When SPTs were performed and interpreted in light of the structured allergy history, 100 (79%) were judged to be clinically allergic to 1 or more of the triggers. Compared with the combined assessment based on structured history and skin tests, patients self-assessments were poor indicators of true allergic status (Fig 1 and Table II). The overall trend was for patients to underestimate their allergic sensitivity, especially for house dust mite and pollens, but they tended to overestimate allergy to domestic pets. In contrast, relying on the structured allergy history alone tended to overestimate patients sensitivity, with false-positive rates between 27% and 75% depending on the allergen being assessed. However, the predictive value of the structured history was comparable with the predictive value of the patients self-assessment, varying from 0.38 for dog to 0.66 for cat (Fig 2 and Table II). One hundred four of the 127 participants returned the questionnaire sent out 3 months after the initial assessment. Almost all of those responding reported finding the allergy assessment and information sheets useful (Table III). The accuracy of recall of allergic status was high (91% for positive results and 82% for negative results, Table IV). Forty-two (40.5%) of 104 said that they had made changes as a result of participating in the study: 35% had made lifestyle changes, and 15% had altered their medications. Eighteen percent reported improvements in their symptoms after the assessment (Table III). The level of agreement between the diagnoses made by primary care clinicians and the allergy specialist was high (Table V). By convention, k values of greater than 0.81 are considered to represent almost complete agreement. 12 In the 9 instances in which discrepancies arose, these all occurred when the primary care clinician allowed data from the skin tests to override a negative clinical history. For all other assessments, there was unanimity. DISCUSSION This study demonstrates the value of SPTs in determining whether allergic triggers are important in a patient s asthma or rhinitis. The combination of SPTs and a structured allergy history is markedly superior to patients own assessments of their allergic status or the structured history alone. At the same time, SPTs should not be used in isolation because reliance on SPTs alone substantially overestimates the clinical relevance of allergic triggers.
J ALLERGY CLIN IMMUNOL VOLUME 123, NUMBER 3 SMITH ET AL 649 TABLE III. Participants views of the allergy assessment as reported in the follow-up questionnaire (n 5 104) Question Response, no. (%) Overall, did you find the allergy test useful? Yes 102 (98.1) No 2 (1.9) If you received an information sheet, was it helpful? Yes 80 (93.0) No 6 (7.0) Not applicable 18 How much discomfort did you experience with the SPT? No discomfort 66 (63.5) A little discomfort 35 (33.7) A lot of discomfort 3 (2.9) Would you have another SPT? Yes 95 (91.3) No 9 (8.7) Did you make any changes to your lifestyle as a result of the assessment? Yes 36 (34.6) No 68 (65.4) Did you make any changes to your medications as a result of the assessment? Yes 16 (15.4) No 88 (84.6) Did your symptoms improve after the assessment? Yes 19 (18.3) No 85 (81.7) Making an accurate causative diagnosis is essential if patients are to be offered targeted advice on avoiding allergenic triggers. In this community-based study population, the patient s own perception of allergic sensitivity was often inaccurate (usually underestimating the role of allergy). This means that patients might not seek out appropriate interventions that could improve their conditions. Conversely, the structured allergy history tended to overemphasize the importance of allergic triggers, and if this was followed blindly to offer allergen avoidance advice, many patients would have received advice to reduce their allergen exposure without any realistic prospect of improvement. The combined approach with a structured history and SPTs is necessary to identify clinically relevant allergic triggers. This approach then has the twin advantages of increasing the chance that allergen avoidance measures will be effective in those to whom they are recommended and reducing inappropriate costs to patients who will not benefit from allergen avoidance. Of course, this does depend on having effective allergenspecific interventions. If drug therapy is cheap, safe, and fully effective, then it could be argued that it does not matter to what patients are sensitized or indeed whether they are sensitized at all. Because antiallergy drugs are cheap and pretty safe, a case can be made for an empiric trial of medication as the initial step for patients with a convincing history of allergic rhinitis or asthma. 13 In patients who are comfortable with taking medication, this pragmatic approach might seem reasonable, but clinicians need to be aware that relying on a convincing history risks misclassifying up to one third of patients, and therefore they need to have a low threshold for rethinking diagnosis and management. The ultimate decision about the preferred management pathway (empiric or test driven) needs to be informed by studies of patient outcome. We have now commenced a trial of management informed by TABLE IV. Accuracy of recall of allergy diagnosis after 3 months Allergen diagnosis Diagnosed (n) Recall of diagnosis (n) Correct Incorrect Accuracy of recall (%) Grass pollen Positive 54 48 6 88.9 (77.8-94.7) Negative 45 38 7 84.4 (71.1-92.2) Tree pollen Positive 43 39 4 90.7 (78.3-96.2) Negative 58 44 14 75.9 (63.4-85.0) Cat Positive 39 38 1 97.4 (86.8-99.4) Negative 62 50 12 80.6 (69.1-88.5) Dog Positive 14 13 1 92.9 (68.1-98.3) Negative 87 73 14 83.9 (74.8-90.1) House dust mite Positive 53 47 6 88.7 (77.4-94.6) Negative 46 40 6 87.0 (74.3-93.8) formal allergy assessment versus empiric treatment in adults with rhinitis and asthma. Ultimately, diagnostic pathways must be validated not by diagnostic precision but by demonstrating that they lead to better symptom control, improved quality of life, and more cost-effective health care. Similar evidence of allergic sensitization could be obtained by using blood tests for specific IgE measurement. However, these tests are relatively expensive compared with SPTs. Some have argued that SPTs are too complicated to implement in primary care, but only limited training is required to perform SPTs reliably, and the technique can be acquired in a couple of hours by a wide range of health professionals. The main limitation of SPTs is the initial outlay for the extracts, but these usually last for about 2 years, and the overall cost per test is small in comparison with blood tests for IgE measurement. Another potential limitation is the time needed to perform SPTs, but this must surely be offset against the administrative time and financial costs of referring patients to specialist clinics or arranging blood tests. Although SPTs add greatly to assessment based on structured history, SPTs are not sufficient in themselves to guide intervention. Within our study group, relying on SPTs alone would have led to inappropriate advice in 32 patients relating to 50 tests (Fig 3). The test results must be combined with the history to achieve an accurate assessment. With this in mind, it is important to emphasize that SPTs are not useful only when the responses are positive. A negative response might help guide the patient and his or her advisers to focus on pharmacologic treatment or, in the case of nasal symptoms, to consider obtaining a surgical review. For this study, a limited panel of airborne allergens was used, reflecting the most common causes of inhalant allergy in the UK. Clearly SPTs with a larger range of allergens might be appropriate in patients who provide a history of multiple sensitizations or whose conditions do not improve despite targeted intervention. Many hospital clinics have larger routine panels, including molds, feathers, horse, and other domestic animals (eg, rabbit, hamster, and guinea pig). Although assessing these sensitivities routinely might help in a few cases, it probably adds little to the basic panel when used in general practice. Of course, patients reporting allergic symptoms to animals other than cats and dogs should undergo specific assessment, but this does not justify the inclusion of extra extracts in routine SPT panels for use in primary care.
650 SMITH ET AL J ALLERGY CLIN IMMUNOL MARCH 2009 TABLE V. Level of agreement between diagnosis made by primary care clinician and diagnosis made by a specialist in allergy Measure of agreement Grass pollen Tree pollen Cat Dog House dust mite Agreement (%) 93 92 97 97 86 Positive agreement (%) 87 80 96 80 94 Negative agreement (%) 86 88 98 96 93 Cohen k 0.854 (0.77-0.95) 0.828 (0.73-0.93) 0.931 (0.87-1.0) 0.864 (0.74-0.99) 0.870 (0.79-0.96) Overall agreement is the proportion of judgments in which the 2 clinicians agree. Specific positive agreement is the proportion of patients given diagnoses of allergy that were identified as being allergic by both clinicians. Specific negative agreement is the converse, the proportion of patients not given diagnoses of allergy, that were identified as not being allergic by both reviewers. in not undertaking costly allergen avoidance measures, which have no chance of being successful. Some inaccuracy of recall was noted after 3 months. On one level, this might not matter if the effect is to avoid cats and dogs when the subject is not clinically sensitive, but retention of the results might be enhanced by providing patients with a written record of the test results and of their interpretation. In conclusion, when used as part of a detailed allergy assessment in primary care, allergen SPTs add precision to the diagnosis of allergy, enabling allergen avoidance to be tailored to the allergic status of the patient. The procedure was fully acceptable to patients and staff. Further work is required to demonstrate improved clinical outcomes when specific diagnosis based on SPTs and a structured history is coupled with appropriate allergen-specific interventions. We thank the clerical and clinical staff at the 4 practices in which this project was based and all the patients who agreed to participate. We also thank Meg Cummins for her assistance in the project coordination and management and Matthew Hankins for his advice on statistical analysis. Clinical implications: The combination of SPTs and a structured allergy history is markedly superior to patients own assessments of their allergic status or structured history alone. FIG 3. Overestimation of the prevalence of allergies if SPTs are used alone compared with formal allergy assessment (structured history plus SPTs). SPTs with standardized allergen extracts are very safe procedures, 9 especially in patients with asthma and rhinitis. A few serious adverse reactions to SPTs were reported in the 1960s, when nonstandardized extracts were in use, but no special precautions are needed for the type of assessment described here. Patient acceptability was high, with very few reporting significant discomfort and more than 91% stating their willingness to undergo repeat testing. When questioned 3 months later, almost all patients were pleased to have undergone the assessment. Almost all subjects found the tests and the information sheets useful, and although only a third changed their lifestyle and one sixth changed their medications as a result, 18% of patients reported improvements in their symptoms after assessment and advice. At first sight, this figure might seem disappointing, but in many cases the assessment confirmed patients previous assessments of allergic status, and therefore no changes in their lifestyle, medications, or symptoms would be expected after the assessment, even if they found the information helpful. For those not found to be allergic to house dust mite, there might be considerable savings REFERENCES 1. Gupta R, Sheikh A, Strachan DP, Anderson HR. Burden of allergic disease in the UK: secondary analyses of national databases. Clin Exp Allergy 2004;34:520-6. 2. The International Study of Asthma and Allergies in Childhood (ISAAC) Steering Committee. Worldwide variation in prevalence of symptoms of asthma, allergic rhinoconjunctivitis and atopic eczema. Lancet 1998;351:12225-32. 3. Ahlstedt S, Murray CS. In vitro diagnosis of allergy: how to interpret IgE antibody results in clinical practice. Prim Care Respir J 2006;15:228-36. 4. Brydon M. The effectiveness of a peripatetic allergy nurse practitioner service in managing atopic allergy in general practice a pilot study. Clin Exp Allergy 1993;23:1037-44. 5. Sibbald B, Barnes G, Durham S. Skin prick testing in general practice: a pilot study. J Adv Nurs 1997;26:537-42. 6. Reid MJ, Lockey RF, Turkeltaub PC, Platts-Mills TAE. Fatalities from immunotherapy and skin testing. J Allergy Clin Immunol 1993;92:6-15. 7. Sheikh A, Levy M. Costs are a barrier to GPs performing skin prick testing [letter]. Br J Gen Pract 1999;49:67. 8. Heinzerling L, Frew AJ, Bindslev-Jensen C, Bousquet J, Canonica GW, van Cauwenberge P, et al. Standard skin prick testing and sensitisation to inhalant allergens across Europe a survey from the GA2LEN network. Allergy 2005;60:1287-300. 9. EAACI Subcommittee on Skin Tests Position Allergen standardisation and skin tests. Allergy 1993;48(suppl 14):48-62. 10. Cohen J. A coefficient of agreement for nominal scales. Educ Psych Meas 1960;20: 37-46. 11. Mackinnon A. A spreadsheet for the calculation of comprehensive statistics for the assessmentofdiagnostic testsandinter-rater agreement. CompBiolMed2000;30:127-34. 12. Landis JR, Koch GG. The measurement of observer error agreement for categorical data. Biometrics 1977;33:159-71. 13. Gendo K, Larson EB. Evidence based diagnostic strategies for evaluating suspected allergic rhinitis. Ann Intern Med 2004;40:278-89.