A cross-sectional study was conducted using a survey distributed to the 65 hospitals of the Catalan public network during the first quarter of 2009 (Appendix 1). The surveys were distributed via the hospital administration departments, who designated a contact person in the Respiratory Medicine departments who could provide additional responses and information if necessary. The results were collected between April and June 2009. A maximum of three attempts were made to obtain the data, with a telephone reminder after the second attempt.

The survey collected information on the activity of the hospital, lung function tests that were performed in each hospital, whether basic (forced spirometry) or complex (lung volumes, carbon monoxide diffusing capacity, arterial blood gases), number of spirometries carried out in the previous year, information on the spirometers, person requesting the spirometries and the location where these are carried out, quality control procedures performed, inclusion in hospital databases, information on the professionals carrying out and interpreting the tests, training received, etc. (Appendix 1).

With regard to the location where the spirometry is carried out, for the purposes of this study, the “lung function laboratory” was defined as a set space identified as such in which at least one nursing professional regularly carried out spirometries, carbon monoxide diffusing capacity (DLCO) determinations and/or lung volumes.

The indicators calculated were as follows: number of spirometries per 100 hospital beds, per 100 discharges, per 100 outpatient consultations, per 100 emergency consultations and per 100 inhabitants. All were calculated for one year. Results were obtained overall and by healthcare region.

A descriptive analysis was initially performed of the survey items for the whole of Catalonia, for each of the healthcare regions and for the healthcare sectors within the Barcelona healthcare region. The number of cases and frequencies of the categorical variables and central tendency statistics and dispersion of continuous variables were calculated. The variation ratio (VR) was used to describe the magnitude of the difference between regions. This is the ratio between the highest and lowest values of the rates for the complete set of regions studied. It is an indication of how much the region with the highest rate varies with respect to the region with the lowest rate.

The number of spirometries carried out in each hospital is distributed using a Poisson law. To explore the association between the number of spirometries carried out in each hospital and the various factors of the survey, a multilevel Poisson regression model was constructed, including the number of discharges from the hospital as the offset. The factors studied were the existence of a Respiratory Medicine department in the hospital, whether the patients admitted were seen by a pulmonologist, the number of lung function tests performed, the availability of a set location for performing spirometries, if there was quality control of the tests, the availability of a data system, formal training for professionals and if they had undergone that training in the last year. An association with a P-value of less than .05 was considered significant. The SPSS v18 programme was used for data management and the STATA v11 for the statistical analysis.

All 65 hospitals in the Catalan public healthcare network completed the survey and it was determined that 8 of them, due to their specific specialist nature, did not carry out spirometries. The results presented below are derived from the 57 hospitals in which spirometries were carried out. The survey was piloted in three hospitals in Barcelona and Vic (in which the authors of the design are currently employed) to assess the feasibility of the items to be completed. Survey completion was the responsibility of the physicians in charge of each of the lung function laboratories, who in turn designated the professional who would complete the survey. We do not have data on their professional category and/or responsibilities.

In 34.5% (n=19) of the hospitals there was a Respiratory Medicine department, in 34.5% (n=19) there was a Respiratory Medicine section, 25.5% (n=14) had a consultant pulmonologist and 5.5% (n=3) did not have any of these three resources. Respiratory patients who were admitted were treated in 24.1% (n=13) of the cases primarily by pulmonologists, in 35.2% of cases (n=19) by internists and in the remaining 40.7% (n=22), care was shared between pulmonologists and internists.

Table 1 shows the type of lung function tests that were carried out in the 57 hospitals in the study. Spirometry was carried out in all centres. In 85.9% (n=49) of the centres, at least three of these tests were carried out. Testing was performed in paediatric patients in all healthcare regions, but only in 67% of the hospitals (n=39). Spirometry was carried out in various locations in or outside the hospital that were not mutually exclusive: 47% (n=27) of the hospitals had a set location; of these, 37% (n=21) had a lung function laboratory as previously defined. Spirometry was also performed in other hospital settings: in 37% (n=21), they were performed in outpatient clinics, in 26% (n=15) in the day hospital, in 10% (n=17) in the conventional ward, in 7% (n=4) in the patient's home, in 14% (n=8) in the Respiratory Medicine ward and in 10% (n=17) in the general medicine ward.

In all hospitals, the spirometer was calibrated, mainly by nursing or technical staff: 81% (n=43) nursing staff, 30% (n=16) technical staff and 17% (n=9) physicians, although in some cases the calibration was performed equally by different professionals. The interpretation of the tests depended on the organization of each centre. In 11% (n=6) of the cases, each doctor interpreted the tests that he/she had requested him/herself. In 87% (46%) of the hospitals, the tests were interpreted by pulmonologists and in 19% (n=10) by the internists as well. In 9% (n=5) of the centres, the results were interpreted by family doctors and in one hospital by the nursing staff.

Seventy-one percent (71%) (n=40) of the hospitals stated that they carried out quality control of the tests. Since the question allowed for an open answer, the centres responded in a varied manner. The most frequent responses were the use of healthy volunteers or professionals, direct supervision of the test or data obtained from the spirometry equipment itself and/or the visual analysis of the flow-volume curves.

In 35% (n=2 0), the information provided by the spirometry was available in the hospital data system: it was entered manually in 9 centres and automatically in 11. In 24.5% (n=13) of the hospitals, the precise contents of the spirometry report were transferred to the hospital database or clinical records. In 20.8% (n=11) of the cases, the spirometry data were not entered in an unbundled format (pdf format), so the data could not be individually manipulated, in only 7.5% (n=4) of the hospitals, the data were entered in an unbundled format, and the spirometry report in its entirety was incorporated in an unbundled format in 3.8% (n=2) of the hospitals.

It was observed that only one hospital selected professionals with previous training, while in the majority of the centres (86.8%, n=46), professionals were selected and then received training. In 43.9% (n=25) of the hospitals, the professionals performing the spirometries received internal non-formal training and in 26.3% (n=15), they received internal formal training. When asked about what type of continuing education was given, 25% (n=14) only carried out training at the time the professionals were selected and 14% (n=8) carried out regular yearly training, while 59.6% (n=34) carried out training spontaneously. When asked how long it was since the professional involved in performing the spirometries had done a course, 14.8% (n=8) of the hospitals answered less than 3 months, 5.6% (n=3) between 3 and 6 months, 20.4% (n=11) less than one year, 24.1% (n=13) between 1 and 2 years and 35.2% (n=19), more than 2 years.

The survey showed that of the professionals requesting spirometry, 57% (n=30) were the hospital pulmonologists, 18% (n=9) hospital internal medicine specialists, 11% (n=6) were other hospital physicians, 3% (n=2) pulmonologists from outside the hospital, 4% (n=2) were primary care physicians and 2% (n=1) doctors from other hospitals.

The rate of spirometries per 100 hospital discharges/year was 14.97, with a VR of 1.82, indicating an 82% greater use of spirometry between the region with the highest rate compared to the region with the lowest rate. The rate of spirometries per 100 inhabitants/year was 1.32, with a VR of 1.53; Fig. 1 shows the wide variability in the Barcelona healthcare region. This region was analyzed by healthcare sector, since it accounts for 67.8% of the whole population of Catalonia (Figs. 1 and 2 and Table 2).

Fig. 1.

Spirometry rates per 100 inhabitants in the Barcelona healthcare region.

(0.73MB).

Fig. 2.

Spirometry rates per centre (Catalonia, 2009). The spirometry rate per centre for the whole of Catalonia and by healthcare region are presented on a logarithmic scale and centred on the mean to improve comparability. Each point represents a hospital centre. Bcn: Barcelona healthcare region; Cat: total Catalonia; Centre: Central healthcare region; Ebre: Terres de l’Ebre healthcare region; Girona: Girona healthcare region; Lleida: Lleida healthcare region; Piri: Alt Pirineu healthcare region; Tarra: Tarragona healthcare region.

(0.07MB).

Table 3 shows the analysis of the association between the numbers of spirometries carried out in each centre and the different factors surveyed, using a multilevel Poisson regression model. Among the various factors investigated, a statistically significant association was observed between the existence of a Respiratory Medicine department at the site and the availability of a set location for performing spirometries. The presence of these factors in a centre increased the spirometry rate by 59% and 56%, respectively. This could explain part (8%) of the difference between the centres, although it remains considerable. In this respect, the median incidence rate ratio (MIRR) is interpreted in such a way that the mean increase in the rate of spirometries between any two centres is 98% in the empty model and 88% in the final model. This suggests wide variability between the rates per centre, as shown in Fig. 1 and Table 2. In this figure, the points represent the spirometry rates in each centre logarithmized and centred at zero for better comparability.

The methodology used in this study was a survey. Although there are no methodological alternatives for conducting the study, the limitations of this information source must be taken into account. In this study, the primary care network was not assessed, but we have recently started a survey of all the public network health centres in Catalonia. Since no other regional, national or international studies have been published, we cannot compare our data with the existing literature. There is no indicator in the literature regarding the number of spirometries per inhabitant; accordingly, one of the objectives initially proposed with respect to the assessment of the variability of the results and a comparison with the literature could not be estimated. For this reason, the statistical and clinical relevance of spirometry rates per inhabitant and year was not evaluated.

Other limitations of the study could include the bias in information derived from the person completing the survey and the fact that the estimated rates could not be standardized for age and sex of the patients. Part of the variability observed could be explained by the different population structure of each healthcare area, as well as severity levels or different comorbidities.