RESPIRATORY MEDICINE (1997) 91, 89-93

The importance of process variables analysis in the

assessment of long-term oxygen therapy by
concentrator

A. GRANADO~, J. ESCARRABILL~,
J. M. BORR&~ AND R. RODRIGUEZ-ROIS~N*

“Agkncia d’Avaluaci6 de Tecnologia Mt?dica (CAHTA), Deparfamenf de Sanifat i Segwefat Social,

Genevalifat de Cafalunya, Barcelona, Spain
tUFISS-Respimt~uia, Sevveide Pneunzologia, Ciufat Sanifciria i Univevsithia de Bellvitge,
l’Hospifalef, Spain
*Servei de Pneumologia i Al~lhgia Respivatbvia, Hospital Clinic, Deparfamenf de Medicina,
Universifat de Barcelona, Spain

The aim of the present study was to evaluate process variables and intermediate outcomes involved in
long-term oxygen therapy (LTOT) by concentrator with the purpose of identifying which of those factors
would be the most influential in the final health outcome of the therapy. A cross-sectional survey was
carried out on a random sample of 111 patients receiving LTOT by concentrator in Catalonia (Spain).
Patients were interviewed and assessed at home by a trained physician, and the variables collected were
arterial oxygen saturation, performance of the concentrators, and patient compliance. Sixty-two patients
participated in the study. Overall, LTOT was appropriately prescribed in 36 patients, of whom only 29
were able to correct their level of hypoxaemia. Patient compliance with treatment was considered
adequate in 19 of those 29 patients. Thus, only 19 of 62 patients (31%) fulfilled those criteria needed to
achieve the expected clinical benefits. Strategies for improving the effectiveness of medical interventions
or technologies ought to consider those factors of the therapeutic process which might influence the
expected health outcomes in a specific health-care context.

RESPIR. MED. (1997) 91, 89-93

Introduction is dependent, in averageclinical practice, on the

Technology assessment activities include the
evaluation of the efficacy, safety, effectiveness
and appropriatenessof emerging and established
health technologies (1,2). Randomized control
trials have proved the efficacy of long-term oxy-
gen therapy (LTOT) as survival rates of patients
with chronic obstructive pulmonary disease
(COPD) and chronic respiratory failure were
shown to increase due to the effect of therapy
(3,4). However, the effectivenessof any therapy
Received 25 December 1995 and accepted in revised form
20 May 1996.
Correspondence should be addressed to: A. Granados,
Agkncia d’Avaluaci6 de Tecnologia Mkdica (CAHTA),
Travessera de les Corts, 13 l-1 59, Pave116 Ave Maria, 08028
Barcelona, Spain.
0954-6111/97/020089+05 $12.00/O
achievement of the different intermediate out-
comes implied in the therapeutic process, along
with its appropriate indication and use. In
LTOT, those process variables are the diagnosis
of chronic hypoxaemia, the correction of the
hypoxaemia which depends on correct prescrip-
tion of oxygen flow rate and proper functioning
of the device, and patient compliance with the
treatment. All those intermediate factors should
be met simultaneously in each treated patient in
order to get the expected health benefit of the
therapy. Therefore, the effectiveness of LTOT
depends initially on the appropriateness of the
therapeutic process.
The aim of the present study was to evaluate
process variables and intermediate outcomes
involved in LTOT by concentrator with the
0 1997 W. B. SAUNDERS COMPANY LTD
90 A. GRANADOS ET AL.
purpose of identifying which of those factors
would be the most influential in the final out-
come of the therapy. The questions include: are
the patients, enrolled in a public financed pro-
gramme, being prescribed oxygen correctly?,
are they complying with its use?, are they
reporting compliance accurately? and, finally, is
the equipment functioning well?
Patients and Methods
In Catalonia (Spain), the total number of
patients included in the public financed LTOT
programme was 3910 in 1991 (prevalence rate
65/100000). Overall, 534 patients (14%) of those
patients were receiving LTOT by concentrator. A
random sample of 111 patients was selectedfrom
the census of all patients who had received this
therapy via concentrator during 1991. A cross-
sectional survey of the 111 patients was carried
out from March to April 1992.Patients who died
during the time interval from the census to the
survey (n=21) or were no longer receiving treat-
ment (n= 14) were excluded from the study. All
patients were assessedat home by the same
trained physician, who was not a member of the
research team. Each patient provided informed
consent before participating in the study.
Data regarding gender and place of residence
of members of the programme were collected
from the Catalan Health Service information
system. A questionnaire was also specifically
designed to obtain the following information:
age, smoking habits, main diagnosis, and type
and hours of concentrator utilization. Level of
percentagesaturation of haemoglobin with oxy-
gen in arterial blood (,SaO,)at rest was measured
using a pulse oxymeter (Ohmeda Biox IVA,
U.K.), firstly whilst breathing room air for
30 min and, secondly, whilst breathing oxygen at
the prescribed flow rate for more than 30 min.
The percentage of oxygen delivered by concen-
trator was measured at a flow rate of 2 1min - ’
when the device was working for more than
1.5min using an Ohmeda 5120. Oxygen Monitor
(U.K.) indicated oxygen concentrations from 0
to 100% (drift range f 1%; monitor linearity
f 1% of full scale).
The criteria established to define the ap-
propriateness of the therapeutic process were:
(1) Diagnosis of chronic hypoxaemia at the
time of” the survey (SaO, I 88% at rest while
breathing room air) (5);
(2) Oxygen flow rate capable of raising
arterial oxygen tension (PaO,) at rest (SaO,
>90% at rest whilst breathing oxygen) (6);
(3) Proper functioning of the concentrator
(percent of oxygen delivered >90% at a flow rate
of 2 1min - ‘) (7); and
(4) Adequate patient compliance.
The latter was measured by a clock fitted to
every device in order to record hours of electric-
ity used for a period of at least 7 days. Adequate
compliance was defined as concentrator utiliza-
tion for at least 80% (2 12 h day - ‘) of the
previously defined appropriate utilization time
(215 hday-l).
Data are expressedas mean f SD. Differences
between self-reported and measured compliance
were examined using the McNemar test. A P
value ~0.05 was considered to be significant.
Results
Overall, 76 patients were considered eligible,
although only 62 participated in the study (82%).
Reasons for non-participation were change of
address (nine patients) and hospitalization at
the time the survey took place (five patients).
The average age of the participant patients was
68 f 9 years, 76% were male, 11% of the sur-
veyed patients were active smokers at the time of
the study, 60% were ex-smokers, and 29% had
never smoked. The most frequent diagnosis
reported was chronic obstructive pulmonary dis-
ease (70%). In 87% of cases, the therapy had
been prescribed by chest physicians.
Inclusion in the LTOT programme was
considered to be appropriate for only 36 of the
62 patients assessedat the time of the survey
(SaO,, 81.0 f 5.8%), while this therapy was
considered to be inappropriate in the remaining
26 patients (SaO,, 92.1 + 2.1%). The latter group
of patients was, therefore, excluded from the
therapeutic process analysis. In addition, 12 of
the 62 devices (19O/,)were found not to work
properly. Defective devices were observed
mainly in the group in which the therapy was
inappropriately prescribed (10 of 12 devices).
Table 1 includes the list of concentrator models
used.
 PROCESS VARIABLES ANALYSIS IN THE ASSESSMENT 0~ LT~T 91

TABLE 1. Concentrator models indication and use of LTOT by concentrator in

the present series.Key factors of the therapeutic
Model Number process which might modify the effectiveness
of the therapy were identified and analysed.
Companion 208 (U.S.A.) 17 Overall, 29 of the 36 patients in whom LTOT
Bunn 3001 (U.S.A.) 16 had been appropriately prescribed were able
DeVilbis (U.S.A.) 13 to correct their level of hypoxaemia (SaO,
Bunn Rx02 (U.S.A.) 9 94.0 ic 2.0%) using the device at the prescribed
Forlip (U.S.A.) 4 flow rate. Of those 29 patients, 19 received
Bunn Lite (U.S.A.) 3 oxygen for more than 12 h day - t as measured
Total 62 by the device’s clock. Therefore, the therapy was
prescribed and used appropriately in only 19 of
the 62 patients.
1 Patients participating in study (n = 62) /

Discussion

The results of this study indicate that LTOT via

oxygen concentrator was prescribed and used
appropriately in only 31% of the patients who
were enrolled in the programme. Therefore, the
overall effectivenessof the therapy might have
been affected by a non-appropriate prescription,
an incorrectly prescribed flow rate, an improper
performance of the device, and/or a low patient
compliance. Some studies have analysed the
influence of all these factors independently
(X-11). However, this is the first time that these
factors have been considered simultaneously in a
Satisfactory Unsatisfactory
19/29 10/29
single sample of patients. The first therapeutic
factor to be considered in the assessmentof the
Appropriate process therapeutic process was the appropriateness of
of LTOT by concentrator
the prescription. Pulse oxymetry does not
19162
replace arterial blood gas measurements in the
FIG. 1. Process variables analysis of long-term indication of LTOT; however, since the error in
oxygen therapy (LTOT) via oxygen concentrator, the accuracy and the precision in most pulse
oxymeters is smaller than 3% (12), it is accept-
with the therapy was reported
Compliance able to consider that patients with a SaO, of
as adequate by 54 patients (87%), although 81.0 i 5.8% have a sufficient level of hypoxaemia
measured compliance was found to be adequate to justify home oxygen therapy. One of the
in only 38 of those patients (67%). The latter indications for home oxygen therapy accepted
group were patients who had used the concen- by Medicare is &‘aO, I 88% (13). The results of
trator for an average of 18.1i 4.0 h day ~ ‘. The the present study showed that LTOT by concen-
group who misreported their compliance (12116) trator were not prescribed in accordance with
had used the therapy for an average of well-established prescription criteria (14) in 42%
8+4&3.4hday-‘. The difference between self- of the study patients.
reported and measured compliance with the A secondfactor that might have influenced the
therapy was significant (P~0.05). appropriatenessof the processwas the capability
Figure 1 summarizes the results of the global of the device to correct the baseline level of
assessmentof the process associated with the hypoxaemia. Only 29 of 36 patients who fulfilled
92 A. GRANADOS ET AL
established criteria for the appropriate prescrip-
tion of the therapy were able to correct their
hypoxaemia levels. Failure to raise PaO, to
therapeutic levels may have been a consequence
of either an incorrectly prescribed oxygen flow
rate (n= 5) or a poor device performance (n=2).
Overall, concentrators did not work properly in
19% of the entire sample, a figure which is
similar to that observed in France (15). Concen-
trators have beenused in Catalonia as an oxygen
source since 1984 (16). At the time of the present
study, there was a regulation from the local
health authorities which aimed to ensure the
periodical control of home oxygen therapy com-
pliance (by checking the concentrator timer), but
it did not explicitly oblige the providers to make
periodical inspections of the device functioning.
The fact that almost 20% of all devices did not
work properly can be used to support the notion
that oxygen concentrators should be equipped
with alarms to detect failures in the delivery of
oxygen concentration. Also, quality control
strategiesto detect these failures must be imple-
mented. The number of patients using concen-
trators in Catalonia has progressively increased.
One year after the present study was undertaken,
23% of the patients on LTOT used this oxygen
source, and the percentage of patients using
concentrators fulfilling the indication for LTOT
was higher than the population of patients with
LTOT supplied by cylinders (17).
The last factor related to the therapeutic
process that should be considered is patient
compliance with therapy. According to differ-
ent published studies, the percentage of non-
complaints can range from 36 to 62% (l&19).
In the present study, measured compliance
(61%) was lower than self-reported compliance
(87%).
The results of the present study showed that
only one out of three patients who were enrolled
in the LTOT programme via concentrator might
have received its expected clinical benefits. In-
appropriate prescription appears to be a key
factor on the inappropriateness of the therapeu-
tic process in Catalonia. Nevertheless, appro-
priateness might also be improved if faulty
concentrators were detected and patient adher-
ence to the therapy was increased. This study
was part of a preliminary effort (20) to re-
organize the prescription of LTOT in Catalonia.
The objective of this work was not to decide
whether LTOT was to be maintained or not; this
should be made by the physician in charge. The
authors’ aim was to generateobjective informa-
tion, regarding the environment, to be used
to elaborate recommendations for clinicians,
health-care providers and administrators.
In patients with underlying chronic condi-
tions, factors related to the therapeutic process
must be re-assessedperiodically if the effective-
ness of the therapy is to be guaranteed (21).
This study showed how therapeutic process
variables, which might be more easily con-
trolled in a randomized control trial but not in
average clinical practice, might distort the true
effect of the therapy and, therefore, modify its
expected effectiveness.The present results indi-
cate that, in the context of the Catalan Health
Service, efforts directed towards improving
adherence to well-designed clinical practice
guidelines for the appropriate prescription of
the therapy might have more effect on the
overall effectiveness than efforts directed to
improve patient compliance alone. The latter
might be the implementation of quality control
measuresto detect faulty devices.
This type of technology assessmentstudy can
provide the objective information needed to
design strategiesto improve the therapeutic pro-
cess and, thereby, the effectivenessof LTOT in
COPD patients.
Acknowledgements
The authors thank J. M. Anto, A. Jovell, and
F. Manresa for their critical comments, C. Vila
for data collection, and A. Boland for edi-
torial help. This study was supported by the
Departament de Sanitat i Seguretat Social,
Generalitat de Catalunya, Spain and the
Comissionat per Universitats i Recerca.
References
1. Fuchs VR, Garber AM. The new technology
assessment. N Engl J &fed 1990; 323: 6735677.
2. Granados A, Borr&s JM. Technology assessment
in Catalonia: integrating economic appraisal. Sot
Sci Med 1994; 38: 1643-1646.

3. Nocturnal Oxygen Therapy Trial Group. Con-
tinuous or nocturnal oxygen therapy in hypox-
emit chronic obstructive lung disease. Ann I&m
Med 1980; 93: 391-398.
4. Report of the Medical Research Council Work-
ing Party. Long-term domiciliary oxygen therapy
in chronic hypoxic car pulmonale complicating
chronic bronchitis and emphysema. Lancet 1981;
1: 681-685.
5. Carlin BW, Clausen JL, Ries AL. The use of
cutaneous oximetry in the prescription of long-
term oxygen therapy. Chest 1988; 94: 239-241.
6. Baodouin SV, Waterhouse JC, Tahtamouni T,
Smith JA, Baxter J, Howard P. Long-term domi-
ciliary oxygen treatment for chronic respiratory
failure reviewed. Thorax 1990; 45: 195-198.
7. ECRI. Oxygen concentrators. Health Devices
1993; 22: 3-24.
8. Booth BM, Kasik JE, Zeman RA, Yeager K,
Lemke SR. Compliance with medical practice
guidelines: The case of home oxygen. QRB 1991;
17: 91-96.
9. Dilworth JP, Higgs CMB, Jones PA, White RJ.
Prescription of oxygen concentrators: adherence
to published guidelines. Thorax 1989; 44: 576-
578.
10. Restrick LJ, Paul EA, Braid GM, Cullinan P,
Moore-Gillon J, Wedzicha JA. Assessment and
follow-up of patients prescribed long term oxy-
gen treatment. Thorax 1993; 48: 708-713.
11 Morrison D, Skwarski K, Barret A, MacNee W.
Domiciliary oxygen therapy in Edinburgh: sur-
vival and compliance with treatment. Thorax
1991; 46: 287P.
PR0c~ss VARIABLES ANALYSIS IN THE ASSESSMENT OF LTOT 93
12. Hannhart B, Heberer JP, Saunier C, Laxenaire
MC. Accuracy and precision of fourteen pulse
oxymeters. Eur Respir J 1991; 4: 114-119.
13. Shigeoka JW, Stults BM. Home oxygen therapy
under medicare. West J Med 1992; 156: 3944.
14. Levi Valensi P, Aubry P, Donner CF, Robert
et al. Recommendations for long term oxygen
therapy (LTOT). Eur Respir J 1989; 2: 160-164.
15. Sous-Commission Technique ANTADIR. Home
controls of a sample of 2414 oxygen concen-
trators. Eur Respir J 1991; 4: 227-231.
16. Escarrabill J, Estops R, Romero P, Manresa F.
El concentrador de oxigeno coma alternativa a la
oxigenoterapia conventional. Med Clin (Bare)
1986; 86: 531-533.
17 Granados A, Escarrabill J, Borras JM, Sanchez V,
Jove11 AJ. Utilization apropiada y efectividad: la
oxigenoterapia cronica domiciliaria en Cataluiia.
Med Clin (Bare) 1996; 106: 251-253.
18. Walshaw MJ, Lim R, Evans CC, Hind CRK.
Factors influencing the compliance of patients
using oxygen concentrators for long-term
home oxygen therapy. Respir Med 1990; 84:
331-333.
19. Vergeret J, Tufion de Lara M, Douvier JJ et al.
Compliance of COPD patients with long term
oxygen therapy. Eur J Respir Dis 1986; 69
(Suppl.): 421-425.
20. Granados A, Escarrabill J, Soler M. Situation de
la oxigenoterapia domiciliaria en Catalufia. Arc/z
Bronconeumol 1992; 28: 264-266.
21. Banta HD, Thacker SB. The case for reassess-
ment of health care technology. JAMA 1990; 264:
235-240.