Development and Content Validity of the New Version of a Patient Classification Instrument 1

The purpose of this study was to reconstruct Perroca’s patient classification instrument; and to assess the content validity of the new version. A group of experts comprising ten nurses appraised the instrument, using the Delphi technique. Data collection took place from March 2008 to February 2009 by means of structured questionnaires sent by electronic mail. The new version contained nine nursing care areas. Agreement levels on the instrument structure amounted to ≥ 90%, and ranged between 80 and 96% in the nursing care areas. The refinement of classification instruments is of great importance to generate valid and reliable data, supporting management decision making on nursing care planning as well as nursing workload measurement.


Introduction
In recent years, the introduction of the evidencebased practice model and the importance of outcomes achievement in nursing services seems to have lead to the increasing use of management tools.In response to this demand, researchers have attempted to develop specific instruments, cross-culturally adapt or improve existing ones, offering users valid and reliable scales.
Validity and reliability are essential criteria to assess the quality of an instrument.The validity concept relates to the extent to which it shows to be appropriate to measure what it is supposed to measure, in other words, the goal for which it is being used.Reliability refers to its degree of precision, that is, when its measures manage to precisely reflect the actual measures of the attribute under analysis (1) .

Patient classification systems (PCS) have been put
in practice in different countries for many decades, with a view to continuous monitoring of the nursing team Rev. Latino-Am.Enfermagem 2011 Jan-Feb; 19(1):58-66.
workload by identifying patients' care complexity.The variations in care needs these instruments capture permit quanti and qualitative staff adjustments, so as to achieve a balanced relation between workload and available human capital.This enhances the quality and cost-effectiveness of care delivery.Knowledge on patients' care needs also allows for effective care and discharge planning, patient distribution across units, care quality assessment and strengthens arguments on additional nursing staff needs (2)(3) .
In Brazil, the Federal Nursing Council (COFEN), patient classification according to Resolution 293/04 (4) , recommends patient classification according to the PCS as one of the variables to support staff decision-making.
It seems, however, that the application of this instrument is being slowly incorporated into daily nursing practice.

The classification instrument
At the end of the 1990's, Perroca (5) developed an instrument to identify patients' nursing care needs and, consequently, measure the team workload.The original instrument (5)(6)  Psychometric property assessment research on this instrument (6) has given evidence of reliability and validity.

Its use at different hospital institutions aroused some
nurses' comments that it did not adequately estimate the patient's care category.To examine this question, the following studies were carried out: comparison with an international instrument (7) , surveying users' opinion on the structure, applicability and reliability of the produced data (8) and analysis of most relevant care indicators inter and intra-care categories (8) .
These research results evidenced a moderate level of agreement among instruments in different care categories (7) .Participating users showed their satisfaction with the proposed classification instrument, except for a certain trend to underestimate the care category the patient belonged to (8) .They also suggested adding other factors that influence nursing teams' workload (8) .
Considering these results, a review of the instrument was started to update its contents and renew its structure, so as to achieve conformity with advances in health sciences and incorporation of new nursing care and management practice trends in the last decade.This paper presents the initial phases of the process to develop and validate the new version of a PCS and aims to: 1. Reconstruct Perroca's patient classification instrument (5)(6) and 2. Assess the content validity of the new version.

Method Research Design
The new version of the instrument was developed in two phases.To produce the items (care areas), the following were taken into account: users' opinion on the original instrument (8) , research results on how the indicators contributed to patient classification in different care categories (8) , literature review and visits to hospital institutions abroad.The reference framework was related to basic concepts of PCS (2,9) , nursing workload (10)(11) , patient care needs (12)(13) , instrument development (14) and analysis of international instruments (3,(15)(16)(17) .Subsequently, content validity, i.e. checking the relevance of the proposed items and their representativeness to adequately capture the concept one intends to measure (1) , was assessed through the application of the Delphi Technique.This technique aims to reach a consensus among experts' opinions on a topic through a series of structured questionnaires, called phases.Answers to each questionnaire are taken into account to formulate subsequent ones.Thus, each phase is constructed on the responses from the previous phase and the process continues until agreement is reached among the participants (consensus) (18) .
The selection criteria to compose the expert panel included a minimum five years of experience in direct patient care and/or nursing management, teaching and research.Hence, ten nurses were included in the panel, five teachers and five working in teaching and care.Experts were female, with a mean age of 48.9±4.9 years, mean professional experience 23.8±8.3years; eight held a Ph.D. and two a Master's degree in nursing.
Regarding their professional trajectory, the predominant work areas were nursing service management, medicalsurgical nursing, intensive care unit, gynecology and obstetrics and pediatrics and neonatology.
answers reached agreement levels of 80% or higher were considered validated.The consensus criterion was previously defined.
The questionnaires, forwarded by electronic mail, comprised a five-point Likert scale, a first part with ten assertions on the instrument structure and a second with seven assertions on each care area.The assertions referred to relevance, clarity, objectivity, care complexity level and applicability of the instrument.Score five was for "I totally agree" and one for "I totally disagree".
When the assertion took the negative form (assertions 5 and 6 in the structure assessment and assertions 4 and 5 in the care area assessment), the score was inverted, according to literature recommendations (1) , which means that one referred to total agreement and five to total disagreement.Experts were also asked to mark whether the care areas should be maintained, excluded or modified.Open questions in the questionnaire allowed participants to transmit other considerations on the structure and care areas.Together with the first questionnaire, the experts were asked to complete data on personal and professional characteristics.

Data analysis
Minitab Statistical Software (MINITAB) version 12.22 was used for statistical analysis.Descriptive data were presented as percentages, means and standard deviation.The Likert scale was considered as an ordinal measurement level and median and quartiles were calculated (Q1 and Q3).Subjective data were grouped, categorized and related according to the study objective.

Instrument reconstruction
In the proposed new version, some care areas from the original instrument were excluded or their contents merged, while others were added to permit more accurate measurement of patients' care complexity and the nursing resources they used.Thus, the following care areas were considered: 1-Care Process Planning To elaborate the score intervals, it was considered that the patient would continue in a given category if obtaining the minimum score for one gradation in all care areas until about 40% of the care areas in the subsequent gradation.

Expert assessment
The Delphi Technique was applied in three phases to reach a consensus on the structure and contents of the new version of the classification instrument.

Delphi Phase 1
Answers to the first questionnaire showed that, regarding the instrument structure, agreement levels exceeded 80%, except for assertion 2 on clear statements (74% -Median -Md 4.0) and assertion 8 on the observation of growing care complexity levels in care area gradations (78% -Md 4.0).The lowest agreement level was for assertions 5 (extension) and 6 (instrument complexity), each corresponding to 38% -Md 2.0 (Table1).Scores range from 1 to 5: the higher the score, the greater the agreement.Md ( Median).Q1, Q3 (Quartiles).
Although the experts agreed on maintaining the care areas, however, suggestions were made to modify the structure and contents of some of them regarding the quantification of activities, ordering of gradations and use of more appropriate terminologies.It was argued that pressure ulcer prevention measures would be better classified as care instead of monitoring, and also that the care area Investigation and Monitoring contained many variables, hampering its assessment.
The experts suggested separating variables in this area in two domains or categories.

Discussion
Refining an instrument is an effortful and long process.The concept under analysis (construct) needs to be clearly defined.Regarding the PCS, the construct comprises the patients' care needs.As they are countless and multidimensional, they cannot be represented as a whole.The goal of the instrument is not to offer a comprehensive list of all care activities nursing performs, as that would make it extensive and exhaustive; instead, the goal is to contain the representative dimensions of care delivery that most influence the nursing team workload.
Thus, structural design aspects, such as the extent and the way items are presented, become fundamental.Efforts were made to make the instrument shorter in comparison with the original version, without compromising the collected information.Literature has recommended some characteristics of classification instruments to enhance their use for decision-making purposes and their acceptability.Simplicity, clarity, objectivity and not consuming much time for completion have been emphasized as criteria (9) .In this study, the experts' evaluation evidenced that it seemed to comply with these requisites.
In this new version of the instrument, the number of care areas was reduced to nine, in comparison with 13 in the original instrument.Some areas were maintained, contents were merged and, also, other areas were added to picture the patients' true needs more reliably.
Many aspects not included in the previous instrument and which consume the nursing team's time were included, such as care planning, help during diagnostic and therapeutic tests, clinical nursing assessment, urgency and emergency care, pressure ulcer prevention measures, among others.
In a previous study (8) , users reported perceiving that the instrument seemed not to show patient complexity as they observed it.This means that the nurses' clinical judgment and objective assessment through the instrument did not seem to be in line.
Hence, in the new structure, besides alterations in the care areas' contents, the score standardization was revised.The previously used method of equal class intervals among the different categories was replaced by different intervals, as patients with more than 40% of characteristics from the subsequent gradation should also belong to the same care category.
Another important alteration was the exclusion of the companion's participation in some care areas (nutrition and hydration, motility, locomotion, personal hygiene, eliminations).Despite acknowledging relatives/ companions' important contribution to patients' recovery, they cannot be obliged to assume the responsibility for the care act, which will always be a nursing responsibility.
As their presence influences the number of care hours, Some studies (19)(20) have attempted to measure the time spent on this activity, with different results.Although accomplished at a teaching hospital, it should be highlighted that different methods were used (estimated and observed time), different process phases were measured, some with a standardized language and computer system, which makes it difficult to compare the findings.
In response to the need to put in practice NCS and cover its influence on nurses' workload, the care area shown that documentation activities represent 13% (21) and 18-27% (22) of the time nurses spend.
Administrative activities were not taken into account, however, for two reasons: nurse supervisors or coordinators are responsible for most of these activities, and not clinical nurses; classification instruments are structured based on the care concept, that is, they measure care delivered when the patient and/or family is present (direct care) and care delivered when they area not present, but which serves to prepare or complete care (indirect care) (23) .Direct and indirect care have represented about 80% of nursing activities, approximately 50% of which refers to indirect care activities (24) .
International literature displays different workload concepts, which consequently influence its measurement form.Some authors consider that it only comprises direct and indirect care delivery (25) , and others that it should also included non-patient related activities, such as unit and nursing team management (11) .
The fundamental goal of classification instruments needs to be highlighted, which is to produce information about the patient for care administration purposes, reflecting the service offered to the patient/family.
Incorporating other variables not specifically related to the patient, such as operating elements at the unit, into the instrument structure is not recommended, as this can compromise data precision (23) .

Conclusion
is based on individual nursing care needs and serves for use with adult patients.It comprises 13 care areas: Mental State and Level of Consciousness, Oxygenation, Vital Signs, Nutrition and Hydration, Motility, Locomotion, Personal Hygiene, Eliminations, Therapy, Health Education, Behavior, Communication and Skin Integrity.Scores for each of these indicators range from 1 (lowest level of nursing care) to 5 (highest level of care complexity).The minimum score is 13 and the maximum 65 points.Through the instrument, patients can be classified in one of the four care categories: Minimal (13-26 points), Intermediary (27-39 points), Semi-Intensive (40-52 points) and Intensive (53-65 points).

3 .
Permits the establishment of a common language among professionals

0 CA1-
-consuming activities that affect nursing care are represented in the care areas.Care Process Planning and Coordination; CA2-Investigation and Monitoring; CA3-Personal hygiene and eliminations; CA4 -Skin Integrity; CA5-Nutrition and hydration; CA6-Locomotion or Activity; CA7 -Therapeutics; CA8 -Emotional support; CA9-Health education.Scores range from 1 to 5: the higher the score, the higher the agreement level.Md ( Median).Q1, Q3 (Quartiles).
This paper discussed the initial phases in the development and validation of the new version of a PCS.The goal was to create an instrument that would permit identifying the quantity of nursing resources patients demand and, consequently, monitor the nursing team's workload.This re-construction was complex, as it involves a range of factors intervening in the nursing work process, which still need more extensive and indepth research.The instrument demonstrated content validity in the experts' opinion.It still needs further testing, however, to assess other psychometric properties, including internal consistency, inter-rater reliability and construct validity.The author hopes the new version can more reliably represent the nursing team's activity and time spent in the care process.

Table 2 -
Expert Opinions on care areas in Delphi Phase 1. São José do Rio Preto, 2009

Table 4 )
. Due to space constraints, the final version of the instrument cannot be presented in this paper, but is available to anyone interested.

Table 4 -
Expert Opinions on the care areas Care Process Planning and Coordination and Skin Integrity.Delphi 3 phases.São José do Rio Preto, 2009