Research

Clinical Validation of Ineffective

Breathing Pattern, Ineffective Airway
Clearance, and Impaired Gas Exchange
JudyCarlson-Catalano, Margaret Lunney, Catherine Paradiso, Joan Bruno, Barbara
Kraynyak Luke, Teri Martin, Margaret Massoni, Susan Pachter
Purpose: To describe the clinical validation of symptoms or defining characteristics of three
respiratory diagnoses. The contributing factors or etiologies of the diagnoses were identified
and the degree of importance of 30 nursing interventions, 75 direct care and 75 teaching, was
rated for each diagnosis and each patient. Threenursingdiagnosesineffective breathingpattern
(IBP), ineffective airway clearance (IAC),and impaired gas exchange (ICE)-were among the
most frequently used, yet no reported clinical studies validated the defining characteristics of
these diagnoses. This study answers the research questions: What are the defining characteristics
of IBP,IAC, and ICE?What are the etiologies of IBF: IAC, and ICE?What are the most important
interventions for IBF: IAC, and ICE?
Design: Standardizedclinical validation using a convenience sample of 76 people hospitalized
with medical and surgical diagnoses, in one U.S. c i and ~ identified as having one of the three
diagnoses. Data were collected in 7 992- 7 993.
Methods: A literature-based concept analysis generated 37 possible defining characteristics
for the three diagnoses which were included in the instrument. The nurse experts conducted a
health history and physical examinationof each patient and decided (a)whether the 37 defining
characteristics were present or absent, (b) the degree of importance of each possible defining
characteristic for making one or more of the diagnoses, (c) the etiologies, and (d) which of the
30 nursing interventions were important for each diagnosis and patient.
Findings: For each diagnosis, many of the 37 possible defining characteristics were judged as
present but few reached the criterion of .50 as important for making one of the diagnoses. Two
of the possible defining characteristicsreached this criterion for IBP, seven for IAC, and two for
ICE. In contrast to the defining characteristics approved by NANDA, the subjective cues of
“expresses fatigue” and “expresses anxiety” werejudgedas important for making one or more
of the diagnoses.
Conclusions: Clinical validation methods allow discriminating among defining characteristics.
Data that are present are not necessarily characteristicof a diagnosis, and the subjective cues
of expresses fatigue or anxiety may be important for making these diagnoses.

IMAGE: JOURNAL OF kHOLARSHIP, 1998; 30~3,243-248.01998SIGMA THETA

NURSING TAUINTERNATIONAL.

[Key words: airway clearance, ineffective;breathing pattern, ineffective;gas exchange, impaired]

* * *

JudyCarlson-Catalano, RN, EdD, CS, FNP, Mu Epsilon & Epsilon Psi, Associate Professor, Radford University, Radford, VA; Margaret Lunney, RN, PhD, CS, Mu
Epsilon, is Professor, College of Staten Island, Staten Island, NY; Catherine Paradiso, RN, MSN, CS, CCRN, Medical Center of Ocean County, Brick Township,
NJ; Joan Bruno, RN, MSN, C, Mu Epsilon, Neonatal Nurse Practitioner, Beth Israel Medical Center, New York, NY; Barbara Kraynyak Luise, RN, EdD, Mu
Epsilon, Assistant Professor, College of Staten Island, Staten Island, NY; Teri Martin, RN, MS, CS, Mu Epsilon, Family Nurse Practitioner, Benedictine Medical
P.C., Hendersonville, NY; Margaret Massoni, RN, MS, Mu Epsilon, Assistant Professor, College of Staten Island, Staten Island, NY; and Susan Pachter, RN, MS,
CCRN, Epsilon Mu, Staff Nurse, Boca Raton Visiting Nurse Association, Boca Raton, FL. This study was conducted by members of the Research Committee of
the Staten Island Nursing Diagnosis Association (SINDA). The authors express deep appreciation to (a) the clinical experts who acted as raters for the study,
Jean Gordon, Faith King, Rita Magnuski, Sarah Newman, Susan Pachter, Catherine Paradiso, Corrine Settlecase, and Marian Smith; (b) the nationally known
experts in respiratory nursing who judged content validity of the instrument and six case studies, Susan Chase, Kaye Greenlee, Janet Larson, and Regina Maibusch;
(c) Arlene Farren for her contributions in the early stages of planning; (d) nurses of the two hospitals for their overall support of the study, especially nurse
executives Margaret Gallagher of St. Vincent’s Medical Center, Nancy Daurio and Josephine Nappi of Maimonides Medical Center; and (e) the College of
Staten Island Department of Nursing for its support of SINDA. Correspondence to Dr. Carlson-Catalano, Radford University, School of Nursing, Box 6964,
Radford, VA 24142. E-mail: jcarlson8runet.edu

Clinical Sidebar: Laurence Parker, PhD, Research Assistant Professor and Director of Health Services and Outcome Research, Thomas JeffersonMedical College,
Department of Radiology, Philadelphia, PA.
Accepted for publication December 3, 1997.
____

Volume 30, Number 3, Third Quarter 1998 /mage:/ournal of Nursing Scholarship 243
Respiratory Diagnoses

urses have a responsibility to continuously develop

knowledge for clinical practice including knowledge for
the diagnosis of human responses (Haughey, 1995).
Knowledge for accurate diagnosis of human responses
is important because it (a) directs nursing interventions,
(b) facilitates the achievement of positive outcomes, (c) ensures Studies in which Defining Characteristics Were Validated
visibility of nursing at a time when the value of nursing services Defining Characteristics IBP IAC . ICE
is being challenged, and (d) enables the incorporation of nurs-
ing data in computer-based patient records. With the emergence 1. Abnormal blood gases 1-3
of the computer-based patient record, nursing diagnoses will be
2. Adventitious breath sounds 1-7
required as the basis of interventions and outcomes. The useful-
3. Air hunger 7
ness of diagnoses in the clinical record, however, depends on con-
tinuous knowledge development through clinical research. 4. Anteroposterior diameter, increased 1-3
Three respiratory responses frequently seen in nursing practice 5. Assumption of a three-point position 1-3
and with all age groups are ineffective breathing pattern (IBP), 6. Bradypnea
ineffective airway clearance (IAC), and impaired gas exchange
7. Chest excursion, altered 2, 3
(IGE) (North American Nursing Diagnosis Association
[NANDA], 1994). 8 Cough 2 1-8
9 Crepitus
10. Dyspned shortness of breath 1-3 2-7
Background 11 Expresses difficulty with sputum 5, 7
12 Expresses feeling of anxiety 7
The diagnoses of IBP, IAC, and IGE were added to NANDA
13. Expresses feeling of pain r t respiratory 3
in 1980 (NANDA, 1994). Since then, research has demonstrated
that these diagnoses are among the most frequent nursing 14. Expresses feeling of fatigue 5
diagnoses made in various settings and with various age groups 15. Expresses perception of chest congestion
(Gordon & Hiltunen, 1995; Hoskins, McFarlane, Rubenfeld, 16. Fremitus, increased or decreased 2
Schreier, & Walsh, 1986; Lutjens, 1993). The high prevalence 17. Hypercapnia
of these diagnoses is expected because people with many varied
18. Hyperventilation 1-3
medical and surgical problems are at risk for these responses.
Other clinical conditions such as immobility also contribute to 19. Hypoventilation
these respiratory problems. Each diagnosis has defining 20. Hypoxemia 1,2
characteristics or cues that help to determine existence of a 21. Hypoxia
diagnosis in a patient. Thirty seven possible defining
22. Inspiratory/expiratory ratio, abnormal
characteristics were identified by the authors through literature-
23. Irritability
based concept analysis, including, but not limited to, nurse
validation studies (see Table 1). The diagnosis of IAC was 24. Mental status, altered 3 7 1-3
studied more often than the other two diagnoses, including 25. Mouth breathing 3 1-3
international studies, because it is frequently used in clinical 26. Nasal flaring 1-3
practice.
27. Orthopnea
In the nurse validation studies reported in Table 1, the
28. Prolonged expiratory phase 1.3 2-7
identification of possible defining characteristics were limited by
reliance on memory of past clinical cases and inconsistencies in 29. Pursed lip breathing
the education and experience of nurses (Wake, Fehring, & 30. Respiratory rate, rhythm, depth
Fadden, 1991). Clark (1994) verified statistically significant 31. Restlessness
differences in ratings of defining characteristics according to
32. Skin color
education and experience. With reliance on memory of past
33. Sputum 1
clinical cases, it is unlikely that data that are present can be
distinguished from data that are important. In contrast, with 34. Tachycardia 113 5 2
clinical validation methods, data are collected at the same time 35. Upper chest breathing
as the human responses (diagnoses) are experienced by subjects 36. Use of accessory muscles
and nurse raters have specific education, experience, and training.
37. Voice sounds
No reported clinical studies of IBP, IAC, and IGE were found.
~

The purposes of our study were to answer the research questions: 1. McDonald I1 985) 2. York (1985) 3. Capuano, Hitchlngs, &Johnson(1990)
What are the defining characteristics of IBP, IAC, and IGE? What 4. Wake, Fehring. & Fadden (1 991) 5. Clark (1994) 6. Boisvert(1995)
are the etiologies of IBP, IAC, and IGE? What are the most 7. Brukwitzki, Holmgren, & Maibusch (19961 8. Malsuki & Otani (19951
important interventions for IBP, IAC, and IGE?

244 /mage:lourna/ of Nursing Scholarship Volume 30, Number 3, Third Quarter 1998

Methods
This clinical validation study followed the standardized methods
described by Fehring (1987) and Carlson-Catalan0 and Lunney (1995).
Three assumptions were: (a) any single human response can co-occur with
other responses, (b) the cues or evidence for a variety of human responses
may be similar, and (c) for these three diagnoses, the clinical judgments
of expert nurses are needed to distinguish cues that are characteristic of
each diagnosis. The study was conducted in two hospitals of two boroughs
of New York City. It was approved by the Institutional Review Boards
of each hospital and informed consent was obtained from patients.
Sample
The population was hospitalized people with medical or surgical
diagnoses who were identified by staff nurses, nurse managers, and study
coordinators as having one of these nursing diagnoses. Some patients were
referred but not accepted for the study because they were not experiencing
any of these diagnoses when assessed. The convenience sample was 76
patients in 2 hospitals, 45 medical and 31 surgical, who were diagnosed
during data collection as experiencing IBP, IAC, or IGE. To be included
in the study, subjects had to be alert and oriented, able to understand
English, and capable of giving informed consent. We determined that a
sample of at least 20-30 subjects was needed for each diagnosis or
variable, hut raters were unable to identify a sufficient number of patients
with IGE who were also able to give informed consent. The mean age of
the sample was 59 years (SD = 19.8). with equal numbers of men and
women. The primary groupings of medical diagnoses were respiratory ( n
= 41), cardiac (n = 9), abdominal post-operative (n = 4). and other ( n =
22) (e.g., dehydration, Hodgkins disease, arthritis, renal surgery). A
majority of subjects (n = 46) reported they were smokers.
instrument
The instrument, Clinical Differentiation of Respiratory Nursing
Diagnoses (CDRND), was developed by the investigators from the
literature-based concept analysis. The CDRND consists of four parts to
record the clinical judgments of raters. Part I was designed for raters to
record “present or absent” for each of the 37 possible defining
characteristics, including those from NANDA. Space was given for raters
to add other defining characteristics as needed. In Part 11, raters recorded
judgments about which of the three diagnoses were highly accurate and
listed one or more etiologies. The criterion for accuracy was that the
diagnosis was “consistent with all of the cues, supported by highly
relevant cues and precise” (Lunney, 1990, p. 16). Cues were defined as
data that have meaning for interpretation of human responses; highly
relevant cues are the defining characteristics of a diagnosis. Etiologies
were defined as contributing factors that precipitate a response. In Part
111, raters selected the importance of each DC on a scale of 1 (not at all
characteristic of the diagnosis) to 5 (very characteristic of the diagnosis).
In Part IV, raters judged the importance of nursing interventions, 15 direct
care and 15 teaching, on a scale of 1 (not at all important for this diagnosis
and this patient) to 5 (very important for this diagnosis and this patient).
These were nursing interventions from numerous literature sources.
Conceptual and operational definitions were developed by the research
team for each of the 37 possible defining characteristics. This was done
through extensive review of the literature and consensual validation among
the research team. Six case studies were developed to train raters and
evaluate the reliability of raters’ use of the instrument.
Four nationally known experts in respiratory nursing judged the content
validity of the CDRND and the case studies. Using the method of item-
objective congruence, content validity was .92 for the CDRND and .93
for the case studies. To estimate test-retest reliability, a random selection
of subjects in the study were reassessed in three hours by the same raters
( n = 10). This time period was selected because the three diagnoses were
considered relatively unstable. To estimate inter-rater reliability, 11
randomly selected subjects were simultaneously assessed by two raters
Volume 30, Number 3, Third Quarter 1998 1mage:lournal of Nursing Scholarship
Respiratory Diagnoses
who separately recorded clinical judgments. Cohen’s Kappa, a
conservative estimate of agreement on diagnoses which takes chance
agreement into account (Fleiss, 1981), was computed. Test-retest
reliability was 1.0 for each of the three diagnoses. This high reliability
reflects the reliability of the tool and the diagnostic ability of raters, but
may also reflect raters’ memories of previous judgments.
Inter-rater reliability was estimated as .63 for IBP and .63 for IAC;
these kappas are interpreted as “good agreement beyond chance” (Fleiss,
1981, p. 218). No subjects with a diagnosis of IGE were included in the
randomized group of subjects assessed by two raters.
Procedures
Eight masters-prepared nurses, including two of the investigators, were
trained as raters to collect data in the two hospitals. The training of raters
for the study took approximately 30 hours of class time, as well as an
unspecified amount of home study. Standardization of training procedures
was facilitated by use of a training manual that included (a) conceptual
descriptions of each diagnosis, (b) information on the concept of nursing
diagnosis and diagnostic reasoning, (c) information about the meaning
of accuracy of nursing diagnosis (Lunney, 1990), (d) the CDRND with
conceptual and operational definitions of the 37 defining characteristics,
and (e) a review of standardized procedures for taking health histories,
conducting physical examinations and making clinical judgments. Of the
six case studies, three were used for training and three were used for
evaluation of reliability at completion of training. At completion of
training, testing of raters established that raters had 100% agreement on
the highest accuracy diagnosisfes) for each case study.
The raters obtained informed consent from subjects before data
collection began. Data collection included a health history and physical
examination. Clinical judgments were recorded including: (a) which of
the 37 possible defining characteristics were present or absent, (b) which
of the three diagnoses were experienced by the patients based on
definitions of the concepts, (c) what were the degrees of importance of
the 37 possible defining characteristics for each diagnosis made with each
patient, (d) what were the etiologies of diagnoses, and (e) what were the
ratings of degree of importance of 30 nursing interventions for each
diagnosis and each patient. Clinical judgments were recorded on Parts I
to IV of the CDRND. Data collection took about 1 hour for each subject.
Data Analysis
Data were analyzed using SAS version 6.10. The defining
characteristics of each diagnosis were determined by computing weighted
means from raters’ judgments in Part 111 of the CDRND. Data from the
five-point likert scale were converted to a 0 to 1 scale which yields
weighted means. With this method, characteristics with weighted means
above S O were accepted as defining characteristics; those with weighted
means above .80 were accepted as major defining characteristics
(Fehring, 1987), which are common criteri used in psychometric studies.
The method of computing weighted means was also used to determine
the importance of nursing interventions for each diagnosis.
Findings
Of the 37 possible defining characteristics identified through
concept analysis, few met the criterion of S O (see Table 2)
when compared with the nurse validation studies in Table 1.
No additional defining characteristicswere identified.Although
a majority of the 37 possible defining characteristics did not
reach the criterion of S O , they were recorded as present for
various percentages of subjects in the study, ranging from 0.9%
(bradypnea, green-colored sputum) to 87% (expresses fatigue).
Some possible defining characteristics that did not meet the
245
Respiratory Diagnoses

Table 2: Defining Characteristics (DCs) of Ineffective Table 4: Etiologies, Frequencies, and Percentages, for Inef-
Breathing Pattern, Ineffective Airway Clearance, and Im- fective Breathing Pattern (IBP), Ineffective Airway Clear-
paired Gas Exchange (n=76) ance (IAC), and Impaired Gas Exchange (ICE)
Etiologies IBP IAC ICE
Ineffective BreathingPattern! Ineffective Airway Clearance Impaired Gas Exchange
I f % f Yo f %
DCs :
wgt. D c s wet. DCS we. Anxiety 8 15 1 2
Means 3 Means Means Arrhythmias 1 2 1 9
Bronchospasm 3 6 2 5
Chest wall compliance, decreased 3 6
Dyspnea .67 Expresses difficulty Abnormal blood gas 36
Cough, weak 1 2 5 12
Expresses fatigue .60 with sputum .71 Expresses fatigue .50 Fatigue 20 38 12 30
Abnormal breath sounds .66 Fear 3 6 1 2 1 9
Expresses chest congestion .59 Fluid accumulation 3 6 1 2
Expresses fatigue .57 Glottis, open 2 5 2 18
Sputum .52 Hyperventilation 4 8 2 5 1 9
Infection/inflamrnation 3 6 4 9 2 18
Cough so lnspiratory muscle strength, decreased 1 2
Expresses anxiety SO Nutrition (obesity) 2 4
Mucociliary transport decreased 4 9 1 9
Note: Weighted (wgt.) means of .50 and above were accepted as DCs. Musculoskeletal impairment 3 6
Obstruction 10 19 4 9 3 27
Pain 9 17 3 7 1 9
Some possible defining characteristics that did not meet the cri- Retained secretions 1 2 4 9 t 9
terion were judged as present in a large percentage of subjects, Sputum, increased volume or viscosity 1 2 13 30
for example, pallor (skin color) was present in 56% of people
in the study. tions for smoking cessation were rated low in importance. This
In the clinical judgment of the expert raters, a majority of the was because raters believed that patients were “too sick” to ad-
patients (55%) experienced either ineffective breathing pattern dress this issue.
or ineffective airway clearance (see Table 3), while the remain-
der were experiencing two or more of the three responses. No Table 5: Degree of Importance of Nursing Interventions
patients were experiencing impaired gas exchange without ei- Determined by Weighted Means, IneffectiveBreathing Pat-
ther IBP or IAC. tern (IBP), Ineffective Airway Clearance (IAC), and Im-
As expected with the variety of medical and surgical prob- paired Gas Exchange (IGE)
lems of subjects, many different etiologies were identified (see
Table 4). Some subjects exhibited more than one etiology. NursingInterventions IBP IAC IGE

The weighted means of a majority of interventions were above Anxiety relief .51 .52 .36
the S O criterion (see Table 5). The raters’ judgments of the im- Breathing techniques .71 .64 .56
portance of interventions to each diagnosis were consistent with Pulmonary hygiene .63 .80 .75
Coughing .49 .75 .67
interventions suggested in literature sources about treatments for
Deep breathing .64 .77 .69
these diagnoses (McCloskey & Bulechek [Iowa Intervention Fluid increase .47 .69 .61
Project], 1996).For example, providing direct care for changes Humidification .38 .54 .39
in breathing techniques was considered more important for Incentive spirometry so .55 .50
people with IBP than those with IAC. Despite the fact that a Inhalation treatment .51 .72 .69
Medication therapy .63 .70 .81
high percentage of the sample consisted of smokers, interven-
Nutrition improvement .53 .54 .69
Pain relief .41 .33 .47
Table 3: Frequenciesof Ineffective Breathing Pattern (IBP), Positioning .63 .61 .75
Smoking cessation .20 .27. .19
Ineffective Airway Clearance (IAC), and Impaired Gas Ex-
Suctioning .08 .19 .11
change (IGE) (n=76) Teaching anxiety relief .56 .56 .53
Teaching breathingtechniques .79 .71 .67
Diagnosis f 70 Cumu- Cumu- Cumu- Cumu- Teaching pulmonary hygiene .67 .80 .69
lative % lative IBP lative IAC lative ICE Teaching coughing .54 .76 .61
Teaching deep breathing .71 .77 .72
I8P only 27 35.5 35.5 27 - - Teaching fluid increase .46 .70 .64
Teaching humidification .36 .50 .42
IAC only 15 19.7 55.2 - 15 - Teaching incentive spirometry .57 .56 .47
49 37 - Teaching inhalation treatment .52 .70 .72
IBP and IAC 22 28.9 84.1
Teaching medication therapy .66 .70 32
IBP and ICE 4 5.3 89.4 53 - 4 Teaching nutrition improvement .55 .57 .75
Teaching pain relief .41 .38 .53
IAC and IGE 4 5.3 94.7 - 41 8 Teaching positioning .72 .64 .72
IBP, IAC, and ICE 4 5.3 100.1 57 45 12 Teaching smoking cessation .22 .32 .19
Teaching suctioning .08 .17 .1 1

246 Image: Journal of Nursing Scholarship Volume 30, Number 3, Third Quarter 1998

Discussion
The prevalence of these responses in subjects with varying
types of medical and surgical diagnoses illustrates the sig-
nificance of these diagnoses to nursing practice. Because
respiratory function is important to health and these re-
sponses occur in people of all ages with acute and chronic
health problems, nurses in many settings should know how
to diagnose and intervene for ineffective breathing pattern,
ineffective airway clearance, and impaired gas exchange.
In this study, we showed that fewer defining characteris-
tics may be needed to make the three diagnoses than those
accepted in studies using nurse validation methods. If fewer
defining characteristics are needed, the efficiency of mak-
ing these diagnoses will be increased. It is more cumber-
some and difficult to identify numerous defining character-
istics than a few.
The literature-based concept analysis identified subjective
cues as possible defining characteristics (see Table 1).
These were not validated in previous studies, however, prob-
ably because subjective cues were not included in the
NANDA-approved defining characteristics. Two subjective
cues that were accepted as defining characteristics in this
study, expresses fatigue and expresses anxiety, are responses
of the whole person when respiratory function is compro-
mised (Paradiso, 1994).
Expresses fatigue met the criterion as a minor defining
characteristics for each of the diagnoses and was rated as
“present” in 86.9% of the sample. The importance of fatigue
to the diagnosis of respiratory function is highlighted by the
findings with IBP, that is, it was one of only two defining
characteristics that met the .SO criterion. Fatigue is associ-
ated with the increased work of breathing and the struggle
to maintain adequate oxygenatign when these responses oc-
cur. Further studies are needed to distinguish circumstances
in which fatigue is a defining characteristic, a co-existing
phenomenon (Tiesinga, Dassen, & Halfens, 1996), an eti-
ology, or a sequelae. As a co-existing phenomenon, it may
be an etiology, that is, fatigue associated with other prob-
lems can lead to IBP, IAC, and IGE.
For ineffective breathing pattern, the finding that the
weighted means of only two of the 37 possible defining
characteristics reached the .SO criterion was surprising. Be-
cause this diagnostic label reflects the physical act of
breathing, we expected that objective cues such as abnor-
mal rate, rhythm, or depth of breathing would reach the cri-
terion as defining characteristics. An explanation of this
may be that dyspnea is a broad concept which incorporates
these objective data (Breslin, 1995). The raters may have
incorporated these objective data with the idea of dyspnea
as a sensation of difficult breathing. Further study of this
diagnosis is needed to substantiate these findings.
For ineffective airway clearance, defining characteristics
that met the S O criterion are logically congruent with the
concept. All of the accepted defining characteristics, except
expresses fatigue and expresses anxiety, are consistent with
the mechanical or physical nature of this phenomenon. The
Volume 30, Number 3, Third Quarter 1998 /mage:/ournal of Nursing Scholarship
Respiratory Diagnoses
subjects’ perceptions pertaining to difficulty with sputum
were considered as more important than the objective evi-
dence of sputum. It may be that a higher percentage of pa-
tients with this diagnosis have difficulty with sputum than
actually have objective evidence of sputum problems.
For impaired gas exchange, the sample size was small and
the subjects did not represent the full range of acuity for
this diagnosis. The requirements for informed consent
meant that patients were not included who were not men-
tally alert. In future studies, researchers should consider the
possibility of obtaining consent from family members so the
sample can fully represent this diagnosis.
These findings may be confounded by the symptoms of
medical and surgical problems. For example, people who
had IAC and also had Chronic Obstructive Lung Disease
were likely to have abnormal breath sounds while people
who had IAC related to immobility may not have had this
abnormality. In future studies, it may be possible to com-
pare two or more sub-samples with different health prob-
lems to describe these relationships.
In 45% of the cases, two or more of the diagnoses were
made which supports McDonald’s (1985) study of the co-
occurrence of the three diagnoses. For IBP, 47.3% of pa-
tients had another diagnosis. For IAC, 65% of patients had
another diagnosis. The diagnosis of IGE was only made
when the other two diagnoses were also made. Possible ex-
planations are that the diagnoses are overlapping, the diag-
noses of IAC and IGE represent progression of IBP, or the
diagnosis of IGE is a sub-category of IBP and IAC. Fur-
ther study, including studies of construct validity (see Re-
search Methods Sidebar) are needed to address this issue.
Differences between this study and previous studies are
probably related to differences in the two methods, nurse
validation versus clinical validation. There should be greater
confidence, however, in findings from clinical validation
methods. The findings suggest that NANDA-approved de-
fining characteristics should be reconsidered for each of
these diagnoses. The sample size of this study is small, how-
ever, and may not adequately represent people who experi-
ence these phenomena. The acuity level of patients with
these problems and the intensity and complexity of data col-
lection preclude large sample sizes in multiple locations,
therefore the findings from many studies using standardized
methods should be combined for knowledge development
(Carlson-Catalan0 & Lunney, 1995). It is imperative that
studies be conducted such as this with other populations to
further validate the defining characteristics of these diag-
noses.
The high percentage of nursing interventions rated as im-
portant for these patients supports NANDA approval of
these responses as ‘‘nursing” diagnoses. A majority of the
nursing interventions met the criterion as important for the
three diagnoses. There were differences in ratings, however,
which indicate that, with time to provide only one or a few
interventions, nurses may select different interventions for
Ieach diagnosis. The perception by clinical experts that a
large number of teaching interventions were important sup-
247
Respiratory Diagnoses
ports the need for teaching expertise at the bedside. In the
future, longitudinal studies are needed with rater follow-up
of selected interventions and the effects on patient outcomes.
This study should be replicated with other populations. The
findings serve as a basis for further research to explore such
questions as: Which interventions are likely to be associated
with positive outcomes for each of the diagnoses? @Zi&
References
Boisvert, C. (1995). Validation of four nursing diagnosesin France:Apreliminary
report. In M J . Rank & P. Lemone (Eds.), CIassifirationof nursing diagnoses:
Proceedings of the eleventh conference(182-188). Glendale, CA: CINAHL.
Bmlh,E.H. (1995). Respiratory responses to unsupported arm lifts paced during
expiration. Western Journal of Nursing Research, 17,91-100.
Brukwitzki, G., Holmgren,C., & Maibusch, R. (1996). Validation of defining
characteristicsof the nursing diagnosis ineffective airway clearance. Nursing
Diagnosis, 7,63-69.
Capuano, T.A., Hitchmgs, KS., & Johnson, S. (1990). Respiratory nursing
diagnoses: Practicing nurses' selection of defining characteristics. Nursing
Diagnosis, 1,169-174.
Carlson-Catalano,J., & Lunney, M. (1995). Quantitative methods for clinical
validation of nursing diagnoses. Clinical Nurse Specialist: Journal for
Advanced Nursing Practice, 9,306-31 1.
Clark,C.M. (1994).Validation of the defining characteristicsof ineffectiveairway
clearance. In R.M. Carroll-Johnson & M. Paquette (Eds.),Classication of
nursing diagnoses: Frowedm ' gs of the tenth conference(334). Philadelphia:
J B . Lippincott.
Fehring,RJ. (1987). Methods to validate nursing diagnosis.Heart & Lung, 16,
625-629.
Fleiss, J.L. (1981). Statistical methods for rates and proportions. New York
John Wiley.
Gordon, M., & Hiltunen, E. (1995). High frequency treatment priority nursing
diagnoses in critical care. Nursing Diagnoses, 6,143-154.
Haughey, B.P. (1995). Science and nursing practice. Clinical Nurse Specialist.
Journal for Advanced Nursing Practice, 9,299.
Hoskins, L.M.,McFarlane, E.A.,Rubenfeld,M.G., Schreier,AM., & Walsh,
M.B. (1986). Nursing diagnosis in the chronically ill. In M E . Hurley (Ed.),
Classification of nursing diagnoses: Proceedingsof the sixth conference.
(319-329). St. Louis: C.V. Mosby.
IowaInterventionProject,McCloskey,J.C.,& Bulechek,GM.(Eds.).(1996).
Nursing intervention classification (NIC) (2nd ed). St. Louis: Mosby.
Lnnney, M. (1990). Accuracy of nursing diagnoses: Concept development.
Nursing Diagnosis, 1.12-17.
Lutjens, L.R. (1993). The nature and use of nursing diagnosis in hospitals.
Nursing Diagnosis, 4,107-1 13.
Matsuki, M., & Otani, E. (1995). Diagnostic content validation for anxiety,
hopelessness, and ineffective airway clearance. In MJ. Ran= & P. Lemone
(Eds.), Classification of nursing diagnoses: Proceedings of the eleventh
conference (275-276). Glendale, CA. CINAHL.
McDonald, B.R. (1985). Validation of three respiratory nursing diagnoses.
Nursing Clinics of North America, 20,697-710.
North American Nursing Diagnosis Association. (1994). NANDA nursing
diagnoses: Definitions and classification 1995-19%. Philadelphia: Author.
Paradjso, C. (1994).Developmentand conceptualdescriptionsof threerespiratory
nursing diagnoses. In classification of nursing diagnoses: Proceedings of
the tenth conference (194). Philadelphia: J B . Lippincott.
Tiesinga,LJ.,Dassen, T.W.N., & Halfens, R J.G. (196). Fatigue: A summary
of the definitions, dimensions, and indicators. Nursing Diagnosis,7,51-62.
Wake, M.M., Fehring, RJ., & Fadden, T. (1991). Multinational validation of
anxiety,hopelessness and ineffective airway clearance. Nursing Diagnosis, 2,
57-65.
York, K. (1985). Clinical validationof two resphtory nursing diagnoses and their
defining characteristics.Nursing Clinics of North America, 20,657-667.
248 Image:Journalof Nursing Scholarship Volume 30, Number 3, Third Quarter 1998
Sidebar by Laurence Parker
This article raises several questions. Why are patients studied who have
more than one diagnosis?Why did the study not focus on sharpening
differentiation among diagnostic categories? What are the signs and
symptoms that occur over time?Why are there no measures of the effects
of differential interventions on differential diagnoses?
While these are all plausible questions, it is important to note the following:
(a) research and development of a diagnostic classification system is a long
process; (b) construct validation is complex, with content validation an
important first stage; and, (c) an initial validation study should be
understood as one step in the process.
A good example of the vicissitudes of the development of a classification
system is the classification of psychiatric diagnoses, which has more than
a 100-year history.' It began to develop in the 1880s with the work oi
Kraepalin, took on a good part of its theoretical structure from
psychoanalysis in the early part of this century, and was finally formalized
into a standard diagnostic system in 1952. In the 1950s and 1960s, after
70 years of development, there were an impressive series of studies
showing poor inter-rater reliability, differential regional patterns oi
diagnosis, and extreme biasing effects in psychiatric diagnosis. The schema
has been through three major revisions since then, which include: (a) the
overlay of a different theoretical perspective, behaviorism; (b) the reworking
of terms such as neurosis and psychosis (fundamental for almost a century
but no longer used in the taxonomy); and (c) the shift from a set of clinical
categories to a multi-axial diagnostic system of five dimensions. When
reading nursing studies, it is helpful to remember that nursing diagnosis i s
only two decades old.
Methodologists usually distinguish among three distinct levels of validation:
content, criterion, and construct.2 Content validity i s achieved by
systematically examining the domain that is being classified or measured.
It usually involves literature surveys and assessment of professional
judgments about the domain. Most professional licensing exams, which
classify competent and incompetent practitioners, are developed solely
through content validation. Criterion-related validity is explored through
the prediction of outcomes or performance. An important validation oi
the SAT test i s its ability to predict college achievement. Construci
validation is the most powerful level. It shows the congruence of a
classification system with a body of theoretical hypotheses. It can involve
demonstrating relations with similar constructs, statistical techniques such
as factor analysis, reliability analyses, convergent and discriminanl
validation, and experimental causal analysis. Construct validation
subsumes and builds on content and criterion-related validation studies.
Because the criterion and construct validity studies are expensive and
elaborate, they are usually not undertaken until many content validity
studies have been performed.
While this research remains a content validation study, it uses real clinical
cases in natural situations. It moves beyond surveys of the literature or
expert opinion. As a naturalistic study, it gives up something in precision
for the generalizabilityof actual cases.' It highlights the appropriate defining
characteristics for three respiratory diagnoses. As part of a process oi
construct validation, studies can now be designed that focus on questions
made salient here, such as whether cooccurring diagnoses provide a useful
basis for differential interventions, or whether the categories should be
further sharpened and differentiated.
Footnotes
'This historyof psychiatric classificationappears in mosttextbooks, eg., H. Cleitman,
1990, Psychology, 3rd ed., New York: W.W. Norton, 750-752.The standard diagnos-
tic system is the Diagnostic and statistical manual of mental disorders, 1952,1968,
1980,1994, Washington, D.C.: American PsychiatricAssociation.
'The validation process is described in texts on psychometrics, e.g., A. Anastasi, 1988,
Psychologicaltesting, 6th ed., New York: Macmillan, pp. 139-164.
)For a discussion of the importance of field studies and generalizability, see L. J.
Cronbach, 1982, Designing evaluations of educational and social programs, San
Francisco: Jossey-Bass, 112-181 (The limited reach of internal validity; Models foi
internal and external inference).