cmbd_ing

THE CMBD REGISTERS

Objectives and general characteristics

The fundamental objective of the CMBD Registers is to create comprehensive, valid databases on the morbidity and activity of both the internment hospitals and ambulatory clinics of the Catalan hospital network. These data must be useful for understanding the pathology being treated, healthcare planning, assessment of resources and purchase of services.

The CMBD Registers compile population-based information on the following healthcare activities:

Acute-care hospitals and specialized psychiatric hospitals (CMBDAH Register):

Conventional hospitalization.

Major ambulatory surgical procedures (CMA).

Day hospitals (HdD).

Social healthcare resources (CMBDSS Register):

Hospitalization in extended-stay care, convalescence and palliative care units Functional Interdisciplinary Social Healthcare Units (UFISS).

Homecare Support Teams (PADES).

Ambulatory care in day hospitals and in diagnostic and follow-up units for cognitive disorders and dementias.

Ambulatory psychiatric care (CMBDCSM Register).

These are a source of accessible information for internal and external users and have to meet the demands made by different fields. In all cases, the release of data is subject to the current law regarding confidentiality and the use of information.

Description of the common variables

The CMBD Registers present coded data according to pre-established criteria, in order to make them compatible and comparable with other information systems.

The following variables are common to all the Registers:

Discharge identification variables:

• Hospital code

• Medical record number

Sociodemographic variables:

• Date of birth

• Sex

• Place of residence

Administrative variables:

• Source of payment

• Date of admission

• Admission status

• Date of discharge

• Discharge status

Clinical variables:

• Diagnoses

The criteria employed in the codification of the sociodemographic and administrative variables are based on the instruction manual for reporting to the respective register. The International Classification of Diseases, 9th Revision, Clinical Modification (ICD·9·CM) was used for the clinical variables.

The CMBDAH Register compiles the healthcare activity data from acute-care hospitals (conventional hospitalization, major ambulatory surgical procedures and day hospitals) and specialized psychiatric hospitals in Catalonia.

Every three months, the hospitals send the data pertaining to their activity to the Register. Once the data are received, they are validated in accordance with the criteria presented in the reporting manual, and each hospital is sent the summary of the validation.

At the end of the year, the acute-care hospitals receive reports on their conventional hospitalization and CMA activity, and how it compares with other hospitals of similar characteristics. Moreover, the data on this type of activity are sent to the Spanish Ministry of Health so that they can be included in the national CMBDAH Register.

In 2002, the CMBDAH Register has received information from 83 acute-care hospitals and 12 specialized psychiatric hospitals (ANNEX 1). Of the acute-care hospitals, 62 belong to the Hospital Network for Public Use (XHUP) while the remaining 21 do not.

The number of hospitals reporting to the CMBDAH has not increased since 2001.

The information contained in the Register represents 100% of the discharges from the XHUP and specialized psychiatric hospitals. As for hospitals not belonging to the XHUP, the discharges reported to the Register reflect 63% of all activity in relation to the information for 2001 provided by the Statistics of Inpatient Health Care Establishments (EESRI).

In relation to 2001, the total activity reported to the CMBDAH in 2002 has risen 2.7% (TABLE 1). These increases have taken place in CMA activity, HdD and specialized psychiatric hospitals. The progressive introduction, in recent years, of ambulatory services for treating processes formerly treated by means of conventional hospitalization explains the increased activity in CMA and HdD and, consequently, decreased hospitalization.

An internal validation is carried out on the data received in order to verify the quality of the information, to detect possible errors in each of the variables and to check their coherence. This validation follows the criteria established in the Instruction Manual for Reporting to the CMBDAH Register. Codification of the clinical variables is done in accordance with the ICD·9·CM, following the Guidelines for the Codification of Clinical Variables of the CMBDAH of Catalonia. These guidelines seek to unify the codification criteria of the healthcare episodes. They obey the criteria established by the national reference group of the Spanish Ministry of Health, and have been drawn up by CatSalut in joint collaboration with the codification group of the Catalan Medical Documentation Society of the Academy of Medical Sciences of Catalonia and the Balearic Islands.

During the internal validation process, two new variables were created on the basis of the original variables: age of the patient (based on date of birth and date of admission) and length of stay (based on date of admission and date of discharge).

It goes on to show the validity of the variables corresponding to conventional hospitalization and CMA in 2002.

Variables identifying the discharge:

• Validity greater or equal to 99.8% for all variables.

Sociodemographic variables:

• Date of birth, age and sex: validity greater than or equal to 99.9%.

• Place of residence: validity of 95.8%.

Administrative variables:

• Validity greater than 99.8% for all variables, except for the transfer hospital variable (reporting is only required when the discharge status is a transfer to an acute-care hospital, to a social healthcare hospital or to a psychiatric hospital), which has a validity of 81.6%.

Clinical variables:

• Main diagnosis: validity of 99.6%.

Average number of diagnoses (main and others) per discharge: 2.41.

• First procedure: reported for 78.3% of the discharges. Where this variable was reported, validity was 99.9%.

Average number of procedures (all procedures) per discharge: 1.37.

• External cause: validity of 79.0% for traumatology-related admissions.

Perinatal variables:

• Validity greater than 95% for all variables.

In general, the validity of the variables rose slightly in relation to 2001.

DESCRIPTIVE ANALYSIS OF THE CMBDAH REGISTER

This report includes the descriptive analysis of the 2002 CMBDAH Register. The information is presented under the following three headings:

• Conventional hospitalization, together with CMA

• Day hospital (HdD)

• Mental health

The data for conventional hospitalization have been jointly analyzed with those of CMA because major surgical procedures generally treated as ambulatory may require admission, depending on the co-morbidity and sociodemographic conditions of the patient or the amount of time needed to reach the hospital. Since one of the objectives of the CMBDAH is to analyze hospital morbidity, a joint analysis must be made.

The data corresponding to day hospitals and specialized psychiatric hospitals are analyzed separately, since they involve types of activity that differ significantly with respect both to each other and to the area of conventional hospitalization and CMA.

The discharge is the unit of measurement and analysis of the CMBDAH Register. Nevertheless, in certain sections of this report where greater emphasis is placed on the person receiving healthcare instead of discharges we will refer to "patients". Information is also compiled and presented regarding all of the discharges of a given patient. Only in this case will we refer to them as "treated people". Thus, despite the fact that both "patients" and "treated people" refer to those receiving care, each term refers to a different analytical level.

These concepts will also be used when we refer to day hospitals (HdD), except as regards the "discharge" concept. In this case, instead of speaking of discharges we will use the term "sessions".

The source of information, when referring to the population of Catalonia in the research presented in this report (population pyramid and hospitalization rate per 1,000 inhabitants), was the 2001 census.

In the tables presenting the age, sex or length of stay variables, discharges are considered to be null and void when any one of these concepts is non-valid.

The remaining variables deemed non-valid have been grouped in the "unknown" category.

In the tables, percentages under 0.05% are presented as "0.0". A complete absence of recorded discharges is given as "-".

The number of treated people is calculated on the basis of the hospital code and the medical record number. The medical record number makes it possible to identify treated people who have received care more than once at the same hospital. However, it should be noted that the data compiled in the Register do not reveal whether a single treated person has received care in different hospitals. This explains why the number of treated people may be slightly overestimated and, consequently, the average number of discharges per treated person underestimated.

The total number of discharges reported to the CMBDAH Register in 2002 was 847,175. These discharges corresponded to 684,191 treated people (TABLE 2). The average number of discharges per treated person (1.24) showed differences with respect to age and sex (FIGURE 1). This indicator rose with age, except in women aged 15 to 44, since obstetrical activity, which does not usually involve readmission, is very significant in this age group. For all age groups, the average number of discharges per treated person was higher in men than in women (1.28 in men and 1.20 in women).

For treated people admitted more than once, the average number of discharges was 2.49. These treated people (16.0% of the total) accounted for 32.2% of all discharges (TABLE 2).

With regard to 2001, there was an increase in discharges per treated person, by group and by sex (TABLE 2). It was seen that even though the total number of discharges rose 0.9%, the number of treated people dropped. The most significant increases involved treated people with more than one discharge.

Age is calculated using two original variables: date of birth and date of admission. Because of the non-validity of one of the two original variables, this age variable was non-valid in 0.1% of the discharges.

The sex variable was non-valid for 0.06% of the discharges.

Of all discharges, 398,950 (47.1%) were men while 447,940 (52.9%) were women. The number of discharges in women has grown annually, both in absolute and relative terms. In relation to 2001, a 1.5% increase was seen in the number of women, while the number of men remained almost the same (0.1% increase).

This circumstance is explained by two factors: the aging of the population and greater comprehensiveness of information in the Register in terms of obstetrical activity.

The average age was 52.2 (51.9 in 2001). This increase was due to the aging of the population and shown by the rise in the number of discharges of patients over the age of 64, 2.2% more than in 2001.

This increase was 5.7% in patients 80 years of age and over.

Differences were seen with regard to sex. The average age of men (52.2) was older than that of women (51.7). However, if obstetrical activity is discounted, the average age of women is 56.3.

Discharges of women from 25 to 39 years of age accounted for 24.0% of female discharges and 12.7% of all discharges.

The age groups accounting for over 50% of male discharges were those of 0 to 19 and 50 to 74 years of age.

FIGURE 2 shows the percentage distribution by age and sex for the discharges of Catalan residents recorded in the 2002 CMBDAH Register, and the Catalan population according to the 2001 census. The hospitalization rate was calculated using these data (TABLE 3).

This rate, calculated on the basis of data reported to the CMBDAH, was 124.8 discharges per 1,000 inhabitants (123.4 in 2001).

By age group, the highest hospitalization rate was that of those over 75, especially those aged 84 and older (397.0). The lowest hospitalization rate of the group was that of those from 10 to 14 years of age (37.8), especially with respect to girls (TABLE 3).

The place of residence variable includes the codes for province, town, district (in the case of residents of the city of Barcelona) and country (in the case of foreigners).

Non-validity for this variable was 12.4%. Errors were largely attributable to lack of information concerning the municipal district of residents of the city of Barcelona. If this factor is discounted, 4.2% is non-valid.

The information provided by this variable makes it possible to establish the territorial distribution of discharges by healthcare region and healthcare sector of residence.

The healthcare region and healthcare sector of residence were assigned to each discharge on the basis of the town code. Assignation was not possible for the discharges of residents of the city of Barcelona, where the sector assignation required knowledge of the municipal district, in view of the fact that this information has yet to be comprehensively recorded. For this reason, all sectors of Barcelona have been treated together.

TABLE 4 shows the number of discharges per healthcare region of residence and the distribution according to the healthcare region of the discharging hospital.

Out of a total of 847,175 discharges, it was possible to assign the healthcare region of residence to 792,909 (93.6%). As for the remaining discharges, 18,601 (2.2%) resided outside Catalonia while the place of residence was unknown for 35,665 discharges (4.2%). However, 30,963 (3.7%) of these discharges only showed the code of one of the Catalan provinces, and so therefore they could be considered residents of Catalonia even though no healthcare region or sector was specified. These discharges have not been included in the analyses presented in this chapter.

Of the 792,909 discharges involving residents of Catalonia, 660,554 (resolution index, 83.3%) were treated in hospitals in their own healthcare region of residence and 132.355 (16.7%) were treated in other regions. Of these discharges, 61,323 (7.7%) were residents of the healthcare regions of Barcelona Ciutat, Costa de Ponent and Barcelonès Nord i Maresme, and were treated in one of these neighboring regions. The remaining 71,032 discharges (9.0%) corresponded to movement between other healthcare regions.

The resolution index for all discharges of patients residing in Catalonia (83.3%) was on a par with those of previous years. Differences in this indicator can be seen for the healthcare region, which oscillated between 71.8% for the Barcelonès Nord i Maresme healthcare region and 93.7% for the Barcelona Ciutat healthcare region (TABLE 4).

As for discharges for whom the place of residence was recorded as being outside Catalonia (TABLE 4), 52.3% were from hospitals of the Barcelona Ciutat healthcare region and 15.3% from hospitals of the Lleida healthcare region, most of whom (2,285 discharges) were residents of the province of Huesca in Aragon. In absolute terms, the hospitals of the Barcelona Ciutat healthcare region produced the highest number of discharges of residents from outside Catalonia. Nevertheless, in relative terms, Lleida was the healthcare region with the highest percentage of residents from outside Catalonia in relation to all hospital discharges (7.2%). For the healthcare region of Barcelona Ciutat, this percentage was 3.1%.

As indicators of the use of hospitalization resources at the territorial level, hospitalization rates have been calculated by 1,000 inhabitants, by healthcare region and by healthcare sector of residence based on the Catalan population according to the 2001 census. These rates have been calculated by the total number of discharges and by those financed by CatSalut TABLE 5).

When assessing these rates, it should be taken into account that the Register records 100% of the discharges financed by CatSalut. As stated in the Quality Control section, coverage was not 100% for discharges not financed by CatSalut, which for the most part take place in private hospitals.

Since the age and sex structure of the territorial units analyzed was not homogeneous, in addition to the crude hospitalization rates, standardized hospitalization rates have also been calculated by age and sex according to the direct method, using the population of Catalonia as the standard (TABLE 5).

With respect to both healthcare region and sector, differences have been seen in the 6 indicators calculated (TABLE 5). Some of the lowest rates have been found where the population is most elderly and vice versa.

TABLE 6 synthesizes the territorial variation of the hospitalization indicators of discharges financed by CatSalut for 2001 and 2002, on the basis of the maximum/minimum rates and the variation coefficient weighed by population. The variation among regions was smaller since, at this level, certain extreme cases in the sector were compensated. Contrary to what one would expect, the variation in standardized rates was greater than in crude rates. FIGURE 3 shows the standardized hospitalization rates of the discharges financed by CatSalut, by healthcare sector of residence. For comparison purposes, it also shows the rate for the whole of Catalonia and the interval of the variation coefficient weighted by population. Of the entire group of sectors, 28 were found within the limits defined by the variation coefficient, 8 plus Barcelona Ciutat which was treated as a single sector were below it, while the remaining 10 sectors were above it.

To evaluate the differences in hospitalization rates per healthcare sector of the discharges financed by CatSalut, one should bear in mind that publicly financed hospital healthcare in a territory is conditioned by the offer of other types of complementary healthcare.

More frequent in urban areas with high socioeconomic levels, the offer of private healthcare coverage by free-choice insurance companies can be accompanied by relatively low public hospitalization rates.

In certain rural areas, where low population density and good communications cannot justify the presence of a hospital, primary healthcare takes on a substitute role in certain aspects of care. In this case, hospitalization rates can be low. From the viewpoint of healthcare policy, the presence of a hospital is justified in areas with low density of population and where communication is difficult, for example, in a mountainous region. In this case, hospitalization rates are higher than average because the offer of care provided by the close proximity of a hospital, however small, increases the demand. The current situation caused by an aging population, in which the proportion of elderly, multiple-pathology patients is on the rise, causes the frequency of admission of these treated people to depend, to a large extent, on the social healthcare resources available in the area.

Source of payment

The source of payment variable records the entity or institution that financed the patient's hospitalization. Of all discharges, 0.13% was non-valid in terms of this variable.

79.8% of all discharges recordedin the Register were financed by CatSalut, a proportion that soars to 93.6% if only XHUP hospital discharges are analyzed. The number of discharges financed by CatSalut in hospitals not belonging to the XHUP and which reported to the CMBDAH Register was 8,650.

FIGURE 4 shows the distribution of discharges not financed by CatSalut by source of payment. The majority of these discharges (74.4%) were financed by free-choice insurance companies, and here there was a 12.1% increase with respect to 2001. Of all privately financed discharges, 23.9% took place in XHUP hospitals while the remaining 76.1% occurred in hospitals not belonging to the public network.

The admission status variable was non-valid in 0.3% of the discharges.

Of all discharges, 51.2% were planned and 48.8% involved emergencies. This distribution in XHUP hospitals was 47.7% and 52.5%, while in hospitals not belonging to the XHUP these percentages were 71.2% and 28.8% respectively. These proportions were the same for 2001.

With regard to the discharge status variable, 0.4% of discharges were non-valid.

TABLE 7 shows the place where patients go once they have been discharged. The most notable increases over 2001 were produced in the discharge circumstances "Transfer to a center for medium or long-term stays" and "Transfer to a social healthcare residence". This circumstance is explained by the high number of discharges involving elderly patients.

As for the number of deaths, an increasing tendency has been seen in recent years, also associated with age. 57.1% of deaths involved patients over age 75.

Discharges following emergency admission showed a higher percentage of deaths than discharges following planned admission, 4.8% and 0.9% respectively. 73.9% of deaths corresponded to emergency medical discharges.

TABLE 7. Distribution of discharges by discharge status. 2001 and 2002

Length of stay is a variable calculated on the basis of the date of admission and the date of discharge. 0.02% of all discharges were non-valid for the calculation of length of stay (this percentage includes stays longer than 6 months).

In 2002, the average length of stay for all discharges was 5.6 days (in 2001, 5.7 days) (FIGURE 5). This decrease can be attributed to a large extent to improved hospital efficiency but increased CMA activity is also a significant contributing factor.

Average length of stay varies a great deal on account of other variables such as age, sex, diagnosis, admission or discharge status, source of payment and so on.

The average length of stay for men was 6.1 days, while for women it was 5.1 days. In general, there was a clear tendency for the average length of stay to increase with age, with the exception of infants of under 1 year and women between 20 and 39 (TABLE 8 and FIGURE 6).

A high volume of hospitalization combined with a long average stay meant that certain age groups accounted for very high percentages of the total number of days spent in hospital. This was the case of those over 70, who accounted for 31.4% of the discharges (TABLE 3) and accumulated 40.0% of all days spent in hospital (TABLE 8). In 2001, these figures were 30.8% of all discharges and 39.5% of all days spent in hospital.

We should also take into account the significance of lengthy stays, because, although stays lasting longer than 30 days represented only 1.9% of discharges, they constituted 16.6% of the total days spent in hospital. In 2001, these figures accounted for 1.9% of all discharges and 16.1% of all days spent in hospital.

If length of stay is analyzed by admission status, it can be seen that emergency admissions were followed by an average stay of 7.6 days whereas planned stays lasted 3.7 days (7.6 and 3.9 days in 2001). The shorter average length of stay for planned discharges was basically due to increased major ambulatory surgical procedure (CMA) activity.

Discharges financed by CatSalut in 2002 were preceded by an average length of stay of 5.8 days, and those not financed by CatSalut, 4.5 days (6.0 days and 4.5 days in 2001, respectively). The shorter average length of stay of discharges not financed by CatSalut was fundamentally due to the lower degree of complexity of the illnesses of the patients.

The average length of stay calculated by discharge status showed noteworthy differences. The higher average length of stay corresponded to patients transferred to centers for medium or long-term stays (14.9 days) or to social healthcare residences (14.2 days) and those for whom the discharge status was recorded as death (11.2 days). In 2001 the average length of stay for these three discharge statuses was 15.5 days, 13.3 days and 11.5 days, respectively.

FIGURE 5. Distribution of discharges by days of stay. 2002

TABLE 8. Days of stay and average length of stay by age group and sex. 2002

FIGURE 6. Average length of stay by age group and sex. 2001 and 2002

The CMBDAH records the main diagnosis of each discharge and up to three further diagnoses, coded according to the ICD·9·CM.

The main diagnosis variable indicates the reason why a patient was admitted to hospital.

The other diagnoses include all illnesses or complications detected in the patient not considered to be the main diagnosis, either already present at the time of admission or developed during the hospital stay.

The percentage of non-valid discharges for the main diagnosis variable was 1.3% of all discharges. Lack of specific information was the cause of non-validity in 0.9% of discharges. The term "non-specific diagnosis" refers to those cases where, although correct, the information failed to present the highest level of detail possible.

In 0.3% of all discharges there was no recorded information concerning the main diagnosis. The variable was either left blank or entered as an unknown diagnosis (code 799.9). The other causes of non-validity were attributable to coding errors not included in the ICD·9·CM and inconsistencies in relation to age and/or sex (0.1%).

54.2% of the discharges provided correct information regarding the first secondary diagnosis, while the percentage for the second one was 39.2% and 27.1% for the third. Non-validity of the other diagnoses was due primarily to lack of specificity, which for the three secondary diagnoses was 0.1%. Non-validity as a result of error or inconsistency was 0.03%.

The average number of recorded diagnoses per discharge (main and other) was 2.41. In 2001 it was 2.38.

The main diagnosis has been analyzed according to two levels of classification:

• Major ICD·9·CM diagnostic groups.

• Diagnostic categories of the Clinical Classifications for Healthcare Policy Research (CCHPR) proposed by the Agency for Healthcare Policy and Research (AHCPR), of the Department of Health and Human Services of the United States.

The ICD·9·CM classifies the more than 13,000 diagnostic codes into 17 major groups, plus code V, which compiles those factors which influence state of health and contacts with healthcare services.

TABLE 9 shows the distribution of discharges by main diagnosis, according to these 18 major groups.

The 5 major groups with the greatest number of discharges were those having to do with the circulatory, digestive and respiratory systems, diseases involving the nervous system and sensory organs, and complications arising during pregnancy, delivery and puerperium. These groups account for 52.4% of all discharges. In 2001 these 5 groups were also the most frequent and accounted for 51,6% of all discharges.

When comparing the number of discharges by major groups for 2001 and 2002, it was seen that 11 of the 18 major groups had shown growth and seven had decreased. The groups with the highest percentage increases were affections originating in the perinatal period and diseases affecting the respiratory system (rises of 12.4% and 6.6%, respectively). The group of endocrine, nutritional, metabolic and immune diseases and the group that included skin and subcutaneous tissue were those showing the sharpest drops (decreases of 4.4% and 4.3%, respectively) (TABLE 9).

Even though each group comprises very different diagnoses and typologies of patients, it should be mentioned that the longest average stays were for mental disorders, affections originating during the perinatal period, and injuries and poisonings. The shortest stays were for diseases of the nervous system and sensory organs (this group includes cataracts which are usually operated by means of CMA) and complications arising from pregnancy, delivery and puerperium.

The diagnostic groups accumulating most days in hospital were diseases of the circulatory system, neoplasms, injuries and poisonings, and diseases of the respiratory system, which accounted for 51.0% of all stays.

Significant differences were seen with regard to the average age. The highest values involved diseases of the circulatory system, nervous system and sensory organs, and neoplasms. The lowest corresponded, predictably, to perinatal and congenital diseases, followed by deliveries and infectious diseases (TABLE 9).

The average age for all discharges increased by 0.3 year. For major diagnostic groups, it should be remarked that there was a rise, in relation to 2001, in diseases of the respiratory system (1.5 years) and in injuries and poisonings (1.3 years).

There were also differences in the distribution of discharges by cause of admission in relation to the age and sex of the treated person (TABLES 10 and 11).

Perinatal affections and respiratory ailments were significant for members of both sexes under the age of 5. Differences were seen with regard to sex and age. What most stood out in women were deliveries (15 to 44 years of age), neoplasms (45 to 54 years of age), diseases related to the locomotive system and connective tissue (55 to 64 years of age), and diseases of the nervous system and sensory organs (65 to 84 years of age). Most notable in men were injuries and cases of poisoning (5 to 34 years of age), diseases of the digestive system (35 to 54 years of age) and diseases of the circulatory system (55 to 84 years of age).

The Clinical Classifications for Healthcare Policy Research (CCHPR) organize all of the diagnostic codes into 259 diagnostic categories, using a clinical criterion above all. These 259 categories are then broken down into 17 groups.

ANNEX 2 sets out the detailed codes that make up each of the diagnostic categories.

TABLE 12 shows, for each diagnostic category, the total number of discharges and percentage distribution, average length of stay, average age, proportion of discharges among women and mortality rate for 2002, as well as the number of discharges and percentage distribution for 2001.

For 2002, it should be pointed out that the categories showing highest growth in absolute terms over 2001 were chronic obstructive lung disease and bronchiectasia (2,146 discharges), acute bronchitis (1,978 discharges), prolonged pregnancy (1,667 discharges), cataracts (1,657 discharges), osteoarthritis (1,226 discharges) and acute myocardial infarction (961 discharges). Those that decreased included chemotherapy and radiation therapy (1,376 discharges), disorders and traumatic dislocations of the articulations (855 discharges), medium otitis and related infections (488 discharges), intracranial injuries (472), infections of the skin and subcutaneous tissue (426 discharges) and stones in the urinary tract (422 discharges).

The following are the most noteworthy characteristics of TABLE 12 for the most relevant categories, compared with those of 2001.

The most noteworthy characteristics of TABLE 12

The CMBDAH Register presents up to 4 diagnostic or therapeutic procedures coded in accordance with the ICD·9·CM.

The first is the diagnostic or therapeutic procedure, which is carried out during admission of the patient and most directly related to the main diagnosis.

This variable was recorded for 78.3% of all discharges, of which 0.2% was not specified and 0.02% was either erroneous or inconsistent with age and/or sex.

As for other procedures, the reported percentages were 38.4% for the second one, 13.9% for the third and 6.6% for the fourth. The average number of procedures recorded per discharge was 1.37 (1.36 in 2001).

As with the main diagnosis, two workings of the first procedure are presented by different group level:

• Major ICD·9·CM procedural groups.

• Procedural categories of the Clinical Classifications for Healthcare Policy Research (CCHPR) proposed by the Agency for Healthcare Policy and Research (AHCPR) of the Department of Health and Human Services of the United States.

The ICD·9·CM classifies the more than 4,000 procedural codes into 16 major groups.

TABLE 13 shows the distribution of discharges by first procedure according to these major groups.

Operations on the musculoskeletal system, digestive system, eye operations and obstetrical procedures accounted for 45.4% of the discharges for which the first procedure was recorded (45.5% in 2001). These procedural groups were always the most frequent, if one discounts the miscellaneous procedural group, which was made up almost entirely of non-surgical procedures.

The number of discharges for the first procedure recorded went up 0.6% with regard to 2001. The procedural groups that increased with regard to this value were: operations on the nervous system, eye operations, obstetrical procedures, and operations on the hematic and lymphatic systems, cardiovascular system and male genital organs.

Even though each group includes very different procedures and typologies of patients, the longest average stays plus discharges corresponded to operations on the hematic and lymphatic systems and the respiratory system, and the shortest to operations on the eyes, ear, nose, mouth and pharynx (TABLE 13).

In terms of age, the procedural groups with the oldest average age were those involving operations on the eyes, urinary tract and cardiovascular system. The lowest average ages involved operations on the nose, mouth and throat, and obstetrical procedures (TABLE 13).

For all discharges with the first procedure reported, the average age increased by 0.4 year. For major procedural groups, it should be pointed out that there was an age increase in relation to 2001, in operations on the skin and teguments (1.2 years), in those of the nose, mouth and pharynx (1.2 years) and in those on the musculoskeletal system (1.1 years).

As in the case of the main diagnosis, there were differences in the distribution of the discharges for the first procedure according to age and sex (TABLES 14 and 15).

In girls aged 1 to 14, the most frequent operations were those involving the nose, mouth and pharynx, while in boys they were those involving the genital organs. Operations on the digestive system were among the three most frequent interventions in men of all ages, especially those from 45 to 74. Eye operations were the most frequent after age 65 in women and age 75 in men. In the 15 to 44 year age group, obstetrical procedures were the most frequent in women, while operations on the musculoskeletal system were for men. This type of operation was the most frequent in women from 45 to 64 years of age.

The CCHPR procedural classification organizes all of the procedural codes into 231 categories and 16 groups.

The level of comprehensive information provided in the procedural statement for the CMBDAH Register varied, according to whether these were surgical or medical. Surgical procedures were comprehensively collected, but medical procedures depended on the criteria of each hospital. While some presented all of them systematically, others did not present any. Because of this varying level of comprehensiveness of information and the fact that certain categories included both surgical and medical procedures, the analysis was presented differentiating between each type of procedure, thus enabling us to determine the exact number of surgical procedures.

ANNEX 3 shows the detailed codes used for each of the procedural categories. Those considered surgical were distinguished from those deemed medical.

TABLE 16 shows the total number of discharges for 2002 for each type of procedure and the percentage distribution, average length of stay, average age, proportion of discharges among women and mortality rate.

Of all discharges, 45.2% were recorded as surgical procedures and 33.1% as medical procedures. No procedure was recorded for 21.7% of discharges. This last group corresponded in large part to non-surgical patients, and for this reason they have been included in TABLE 19 under medical procedures.

The most relevant differences between the two procedural groups, surgical and medical, had to do with the average length of stay and the mortality rate. The average length of stay of discharges with surgical procedures was shorter because, first of all, this group included CMA activity and, second, a large number had to do with non-complex pathologies. The mortality rate was notably higher for discharges in the medical procedural group.

In 2002 it should be stressed that, of surgical procedural categories, those that most grew in absolute terms with respect to 2001 were operations on the crystalline lens (1,438 discharges), percutaneous transluminal coronary angioplasty (865 discharges), caesarians (636),