Cancer patients' needs during hospitalisation: a quantitative and qualitative study

Background The evaluation of cancer patients needs, especially during that delicate period when they are hospitalized, allows the identification of those areas of care that require to be improved. Aims of the study were to evaluate the needs in cancer inpatients and to improve the understanding of the meanings of the needs expressed. Methods The study was conducted during a "sample day", with all the cancer patients involved having been hospitalized at the Istituto Nazionale Tumori of Milan (INT) for at least 48 hours beforehand. The study was carried out using quantitative and qualitative methodologies. The quantitative part of the study consisted in making use of the Needs Evaluation Questionnaire (NEQ), a standardized questionnaire administered by the INT Psychology Unit members, supported by a group of volunteers from the Milan section of the Italian League Against Cancer. The aim of the qualitative part of the study, by semi-structured interviews conducted with a small sample of 8 hospitalized patients, was to improve our understanding of the meanings, implications of the needs directly described from the point of view of the patients. Such an approach determines the reasons and conditions of the dissatisfaction in the patient, and provides additional information for the planning of improvement interventions. Results Of the 224 eligible patients, 182 (81%) completed the questionnaire. Four of the top five needs expressed by 40% or more of the responders concerned information needs (diagnosis, future conditions, dialogue with doctors, economic-insurance solutions related to the disease). Only one of the 5 was concerned with improved "hotel" services (bathrooms, meals, cleanliness). Qualitative analysis showed that the most expressed need (to receive more information on their future conditions) has the meaning to know how their future life will be affected more than to know his/her actual prognosis. Conclusions Some of the needs which emerged from this investigation could be immediately satisfied (the need for psychological support, the need for economic aid, the need for spiritual support), while others will have to be faced in the longer term; for example, the presence of a high percentage of needs in patient-physician relationships and/or information-communication issues, could be resolved by setting up structured introductory training courses for all clinicians in the institution. On the other hand, the needs related to the living infrastructure (bathrooms, meals, etc...) could encourage the Institution to improve its services.

support), while others will have to be faced in the longer term; for example, the presence of a high percentage of needs in patient-physician relationships and/or information-communication issues, could be resolved by setting up structured introductory training courses for all clinicians in the institution. On the other hand, the needs related to the living infrastructure (bathrooms, meals, etc...) could encourage the Institution to improve its services.

Background
Over recent years an important change has taken place regarding the caring of patients. The biomedical model characterized by an objective approach to patients, which considers both their bodies and their diseases as objects, has been replaced by the "bio-psycho-social" model. This model is characterized by an approach that considers patients holistically and their disease as a complex event consisting in an alteration of the system at several different but integrated levels, biological, psychological and social [1][2][3].
This transformation in the field of medicine has taken place, above all, with regard to chronic and degenerative diseases such as cancer [4]. Facing a long course of a disease that is often susceptible to control but never to a complete recovery, evaluation of the "the best possible treatment" concerns aspects which are not only necessarily or strictly clinical (undesirable effects of treatment, years of survival). These further aspects are directly linked to the patients' quality of life, to personal aspirations, values, and quality of their relations and needs. For clinicians, examination of these aspects means that they need to pay more attention to the patients themselves as individuals during the various stages of their disease, beginning from diagnosis until the terminal phase. [5,6].
Helping patients to make important decisions about their lives and their health means therefore helping them to be aware of the impact that different treatments may have upon their ability to look after their own families or continue to work [7][8][9][10].
Thus, in the caring of patients, an important task of the physician consists in informing him/herself and informing patients about what can be offered them in order to improve their independence and quality of life [7]. In order to reach this aim, the considerable amount of work that has been carried out on the development of instruments self-compiled by the patient, is in its own way an improvement upon the limits of physician centered evaluation of the physical, psychological and social problems of the patient. Doctors tended to attribute to these problems with their own personal hierarchy of values and priorities, which were, therefore, often truly different from the patients' own perceptions of the problems [11][12][13][14][15]. Furthermore, it has been observed that it is the patients' subjective evaluation of their illness, rather than scientific achievements in medicine, that play a major factor in determining their choices and decisions [8,16]. Added to this has been the patients' difficulty in expressing their own needs. Personality, cultural and historical differences may make mutual comprehension between patient and doctor very difficult.
Until recently, the effort made to bring together these two "realities" (the world of the doctor and the world of the patient) has resulted in an exponential increase in the number of instruments available. These instruments were created to evaluate quality of life, needs and satisfaction of care. A recent review of 210 published articles identified 18 potential need assessment standardized instruments used in cancer research [17].
A careful attention was given to improve psychometric proprieties of these instruments (reliability, validity, responsiveness). Less attention, on the other hand, was given to understand the deep meaning of what patients express when they report aspects of their quality of life, satisfaction and needs [18].
This kind of in-depth studies, which include an analysis of the patients' language and of the connotations and meanings of their words, may well represent a further approach to understanding the patients' conditions and their necessities, and may act as a guide for all those who administer care. The interest in patients' needs, and the meanings that patients give them, led to the setting up of a qualitative study parallel to a quantitative one on psychosocial and assistance-related needs of hospitalized cancer patients, with the aim of inducing interventions that satisfy the majority of their present needs.
The aim of the study was to evaluate different needs' prevalence in cancer inpatients and to understand their deeper meanings in order to promote an improvement in the quality of care. Secondary aims of the study were: a) to confirm the structural validity and the internal consistency of the standardized questionnaire used; b) to evaluate the association between the needs expressed and the basic demographic and clinical characteristics of the sample.

Patients
The survey was conducted in most clinical units at the Istituto Nazionale Tumori (INT) of Milan. The Units excluded were pediatric (as the questionnaire is to be filled out by adults) intensive therapy and nuclear medicine (as physical condition of the patients of such patients does not permit an easy interaction), medical and surgical dayhospital (as patients are admitted for too short a time). All patients admitted to the hospital at least 48 hours before, were invited to participate in the study and asked to give oral (informed) consent. Data regarding sex, age, educational level, primary tumor site, number of previous admissions and days of hospitalization were gathered. The NEQ (Needs Evaluation Questionnaire) was used for needs assessment [19].

The Questionnaire
The NEQ is a standardized questionnaire, consisting of 23 items, developed and validated at the Psychology Unit of INT (see table 2 for a detailed itemized presentation) [19]. It asks about different types of needs: information concerning diagnosis/prognosis (items 1 and 2), information concerning exams and treatment (items 3 and 4), commu-nicative (items 5 -8), and relational (items 20 -22), which make up four distinct factors; twelve further single item scales related to assistance and treatment (items from 9 to 13), to structure (item 14), to financial aspects (items 15,16) and to support (items 17,18,19,23) complete the questionnaire. The NEQ was administered by the members of the Psychology Unit, supported by volunteers from the Milan section of the Italian League Against Cancer, who also assisted the patients in filling in the questionnaire when help was requested.

Statistical Analysis
We compared groups of responders and non-responders with regard to demographic and clinical characteristics, using Fischer 's exact test and Student's t-test.
Confirmatory factor analysis on thetracoric correlation among the observed variables was used to test the structure of the questionnaire described above. Estimation of the model parameters was obtained by means of weighted least squares (WLS) using the program LISREL 8.0 [20]. Evaluation of the model to be confirmed was based on the following criteria: the Chi-square statistic (χ 2 ) [21], which indicates the lack of fit of the model (contrary to conven- tional significance paradigms, large values of chi-square, and small P-values, indicate poor fit and model performance, leading to the rejection of the hypothesis that the assumed model is 'true'); the Adjusted Goodness of Fit Index (AGFI) [22], which, contrary to χ 2 does not depend on sample size and measures how much better the model fits compared to no model at all (it should lie between 0 and 1, large values being associated to good models); the range of the Squared multiple Correlations (SMC) [20,23] or each of the observed variables in the model (each index shows how well an observed variable serves as measurement instrument for the latent variable, with large values (range 0-1) being associated with good fit). Cronbach's alpha indexes [23] were calculated to estimate the internal consistency of the four factors tested.
Association between each of the 23 items of NEQ and sex, age, primary tumor site, educational level, number of previous admissions, self-sufficiency in movement and hospitalization duration was examined by means of Fischer's exact test for rxc tables or Student's "t" test. Multiple comparison influence on Type I error was overcome by splitting the sample into two of equal size: the first, the "training" sample, was used to perform association analysis with an explorative aim and where p < 0.10 was considered significant; the second, the "validation" sample, was used to replicate the analysis and verify whether an as-sociation from the training sample still showed significance.

Methods of the qualitative study
The qualitative part of the study was aimed at complementing the quantitative description of the patients' needs, showing the multiplicity and richness of meaning that the affirmations of the questionnaire had in the subjective world of the individual patient. According with this aim, among the patients who filled out the questionnaire (NEQ), we choose a purposeful sub-sample (eight patients) identifying those who answered YES to more than five items and among them those who demonstrated great skill in recognising and expressing the experiences relative to their condition as admitted patient.
This kind of sampling focuses more on the identification of information-rich cases than on the representativeness [24]. The objective of the study is not to generalise the results but to show and document the complexity of an experience.
By means of individual semi-structured interview, the selected patients were invited to examine the meaning of the need they expressed through the questionnaire. The interview was constructed on the bases of the items of the NEQ questionnaire choosing those considered less deepened in the scientific literature or those who may lead to multiple interpretations by the patients.
The items were the following: n. 2, 5, 7, 11, 13, 17, 18 (the text of the items can be found in table 2). The interview started with the identification of a critical incident regarding the specific situation where the patient felt their needs been unsatisfied and, after this, the subjective meaning of the situation as well as the different implications of the needs were analysed and discussed. The interviews were tape-recorded and later transcribed.
Data analysis was carried out autonomously by three researchers, by means of successive steps, each of which represents an increasing level of generalisation: • Each researcher identified all propositions held personally to be significant, without considering their relations with other parts of the text, accompanying them with his own comments.
• The comments were developed and contextualised along with the entire interview.
• From each interview the researcher extrapolated the themes he/she felt were relevant.
• Analyses of each interview were confronted with the others, with the aim of finding common themes and verifying conformities in similarity or contrast.
• The three analytic processes were confronted with one another.
The results were presented following the NEQ major factor groups obtained in the previous validation: 1. Information 1st factor: refers to information relative to diagnosis and future conditions.
2. Information 2nd factor: refers to information on therapeutic choices and on results of tests.
3. Communication: predominantly regards relations with clinicians, in particular the feeling of being (or not) in a reassuring and sincere relation, based on dialogue.
4. Relational factor: regards the perception of patient's own position in a wider context (family, other patients, society in general).
5. The area of "intimacy" was also flanked beside these four factors.

Results of the quantitative study
Of the 224 eligible patients contacted in the sample day, 182 (81%) fully completed the questionnaire, 2 partially completed it (more than 3 items missing), and 40 did not completed it at all. The reasons for non-completion were the following: refusal to fill out the questionnaire (8 patients), physical and/or cognitive difficulties (12 patients), absence of the patient from his/her ward for surgery or other exams (17 patients), other organizational problems (3 patients). The questionnaire was filled in by 49% of the patients unassisted, while the remaining 51% asked for assistance (the interviewer read the questions to the patient and then wrote down the answers, without exerting any influence on the responder). Table 1 shows the basic demographic and clinical characteristics of the sample of 182 responders: 90 were males and 92 females, with a median age of 60 years and a medium-low educational level (59%). 41% of them were at their first admittance in hospital, while 34% had had 2 admissions or more, with a median duration of the current hospitalization of 7 days. The most common primary tumor sites were: head and neck (19%), colon-rectum (16%) and breast (11%). 91% of the patients were self-sufficient in movement. No statistically significant differences emerged between the group of responders and non-responders as regards gender, age, primary tumor site and duration of hospitalization, while they differed in number of previous admissions: responders had a mean number of 1.5 previous admissions vs 0.7 of non responders (p = 0.018). Table 2 presents the percentage frequency of positive answers and missing values for each item of the questionnaire. The percentages of missing values for each item are very low (range 0-3%) indicating a good comprehensibility and acceptability of the questionnaire by the responders. Among the five requests most frequently expressed by the patients, four regard information needs: concerning diagnosis (40%), about future conditions (61%), regarding a better dialogue with clinicians (45%), and about economic-insurance information (40%), while only one regarded the need for better services at the hospital (bathrooms, meals, cleaning) (59%).
Most of the support-assistance items remain under 30%. The need that was less frequently expressed concerned help for eating, dressing, and visiting the bathroom (11%), which is in accordance with the characteristics of a very low percentage of the sample presenting problems on movement. There was also a low level of request for better attention from nurses (13%).
The confirmatory construct analysis showed that the model tested (on the four factors described in the methods) provides a good fit to the data (Pχ 2 = 0.136, AGFI = 0.98, and range SMCs: 0.61-0.99). Table 3 presents the standardized Lisrel estimated for the factor loading of the model tested, all of which are significantly different from zero at P < 0.001. Cronbach's alpha indices for the four factors tested were 0.63, 0.76, 0.76, 0.64 respectively, showing acceptable to good levels of internal consistency.
None of the association which proved statistically significant on the training sample was confirmed by the analysis on the validation sample. Since the power of tests performed may have been greatly limited by the splitting of the sample for the training and testing procedure, some of the associations thought not statistically significant which maintain their direction in both samples ought to be verified in future studies. The above associations indicate that, as the number of previous admissions grows, the level of information needs (all items except for better dialogue with doctors and better attention from nurses) diminishes; further, that the higher the educational level, the lower the need for economic help.

Results of the qualitative study
The original expressions of the interviewees are reported in tables 4,5,6,7,8,9, grouped into thematic areas corresponding to the principal factors of NEQ. Table 4 shows the need to obtain a diverse modality of communication between clinician and patient, and to redefine the setting in which this traditionally takes place. In addition, table 5 shows the need for making the scenario of a patients' future life intelligible and foreseeable; patients ask for personalized explanations shaped as much as possible to their individual needs, thus avoiding inde-terminacy or generalization based on statistical reference or analysis. According to table 6 clinicians are requested to understand the patients' demands even though these demands are not always consonant with the clinician's own perspective on therapeutic choices. A less hurried, hermetic, and prudent, and a more sincere and polite form of communication was also requested (table 7). Some patients brought to attention the potential danger of communication between hospitalized patients themselves, through which anxiety might be exacerbated among "the deaf", rather than any reassuring exchange of experiences and reciprocal encouragement occurring (table 8). The patients' reflections on the theme of intimacy (table 9) considered in all its multiple of facets -refer, among other things, to criticism of situations like the "doctors' rounds", perceived as a violation of the some of the most elementary principles of respect for privacy.

Discussion
The analysis of the structural validity and the internal consistency of NEQ confirmed the psychometric properties shown in our previous study giving a further evidence of its value in assessing hospitalized patients' needs.
This study shows that some needs are unsatisfied in one out of two hospitalized cancer patients. Most of the needs concern the request to receive more information both about their future conditions (this is the most expressed need, reported by 6 patients out of 10) and their diagnosis (4 out of 10). These findings agree with data already reported [17,[25][26][27]. Besides the documented presence of this two needs and the importance of information in the doctor-patient relationship, some studies have underlined the necessity that clinicians help patients to understand information [25,28,29].

Quality of Life
By 'future' I mean also the quality of life because someone can say "you will live for 10 years", but "how will I live?". More detailed information is needed; like, for example: "you must cope with this or that"; but doctors just give summarized information. This is true above all about oncologists. Also I mean the possibility of being cured, but here everything is reported in statistical terms; it is very schematic: "50% responded like that; 30% like this". But no one knows what their own percentage will be. The patients' desire to know more about their own diagnosis and future conditions does not necessarily mean that the patients ask for an active role in medical decisionmaking. Less than a third of patients expressed the request for deeper involvement in therapeutic choices. Two recent studies on women with breast cancer gave similar results; about half of the women had a role in decision-making that matched their desired role, a quarter had a less active role than desired, and a quarter had a more active role than desired [30,31].
The new trend that encourages self-determination and responsibility by the patients did not, in fact, take into account patients' disorientation, and therefore their inability to take at times a more active role in the management of their experience with the disease. The clinicians find themselves lacking guidelines for managing doctor-   patient relationships. The traditional paternalistic doctorpatient relationship is no longer acceptable; however, the basis of a new clearer and "shared" model has yet to be found.

Communication with the patient is hurried and fairly hermetic also for a person of average cultural background. Therefore given that this is an illness which has important psychological implications, communication in 'snatches' will not do. Communications are given in a very formal context (the 'round'), but it is obvious that for the patient it is more difficult to ask for explanations from 10 doctors instead of just one. They give the impression of speaking in abbreviations. They speak to one another about the results of tests in such a condensed way that it almost gives the impression they don't want to communicate to the patient at all; that the information is more or less for their own internal use. Sometimes they give contradictory information among themselves (depending on the doctor) and this worsens the situation. One time I said to one of them that because patients have so little time available (with the doctors) their 5 minutes of explanation in the morning is meditated on later on in the day
Clinicians should, therefore, consider the meanings (explicit and implicit) contained in a request for more information. The analysis of qualitative data of the study shows that the need for more information on their future conditions does not mean for patients "knowing all there is to  know about prognosis". The patients' principal need appears to be receiving more information about the impact of their illness and treatment on their daily lives, with particular attention being paid to the realistic ability to carry out such activities; in brief "what will I be able to do or not do?", "what will I or won't I be able to eat?", "will I be able to move, work, love, have children and if "yes", how?". One important dimension, therefore, regards the quality and not just the quantity of life remaining. The appropriate care for patients will be the one that also considers the implications of the illness, and is capable of providing patients with assistance for the diverse problems [7].
Another aspect which emerges from the analysis of qualitative data concerning information on prognosis, regards the quantification of the risks of the therapy proposed by the doctor. In the quantification of the uncertainty of evolution and efficacy of a cure, the epidemiological and statistical tools might precisely and accurately render the expectations for a patient with similar characteristics without being able to give a deterministic estimation in specific cases. On the other hand it appears that it is more important for the patient to a) receive personalized information, which is commensurate with his/her physical and emotional situation, and b) to be considered within his/ her own unique profile. Thus "expected time" can be different from "lived time", where the days and nights are diverse, some terrible, others welcoming, others without identity but all intrinsically offering some quality of life.
Furthermore, patients' observations suggest the necessity of an improvement regarding the circumstances under which personal information is discussed. Crowded, noisy ambulatory conditions too often provide the stage for delicate and important communications. Patients respond negatively also to "the doctor's round" on the ward; they are aware of being the object of the clinicians' deliberations and they feel inhibited about open communication.
The request for more dialogue, expressed by around half of the patients, emphasizes the importance of communication and listening skills necessary for selecting information considered relevant by patients. The clinician-patient relationship should, therefore, encompass and integrate listening more attentively to these expressed needs, and the provision of information. It should be a connection between the world of cancer as a 'biological entity' (disease) and the world of cancer as 'a lived experience' (illness). It is within this kind of relationship that the patients will find greater assurance. It will allow the patient to feel that the doctor is nearer to him/her about any future eventuality.
The need for reassurance is expressed by 4 patients out of 10, together with the desire to talk to people who have had similar experiences. Relations among hospitalized patients, however, are complex. Analysis of our qualitative data shows that the relationships between such patients cannot always be simply defined as "having company" or "sharing similar experiences"; it can also represent a moment of emotional engagement. Confrontation with others may sometimes generate an element of 'contagion'. Repartitioning of physical space in hospitals should, therefore, take into consideration the necessity for the setting up in some cases of protected areas in which the patient can expect intimacy.
The need for a more respect of intimacy (present in only 2 patients out of 10) appears to be not only manifested by patients being seen naked, or by their forced sharing of bathrooms. The patient is usually subordinated and in a weak position with regard to the rules and regulation of hospitalization that, at times, fails to respect their intimacy and privacy.
One last, but by no means least, aspect is the satisfaction of needs regarding services offered by the institution, both concerning assistance (control of symptoms and/or of practical help with functional problems) and hotel services (bathrooms, meals and cleanliness). While there seems to be little presence of needs concerning assistance and most patients are satisfied with hospital staff, the needs concerning hotel services were expressed as being not satisfactory by two patients out of three, indicating the need to improve these services.
The 'humanization movement' in hospitals, which in recent years is definitively culturally based, will be accomplished not only through education and training processes of health workers, but also by structural interventions aimed at rendering the experience of hospitalization more amenable and comfortable for patients [32].
It is difficult to evaluate the real impact of this research to meet the needs of the patients of our hospital. Preliminary data from this study were presented at a meeting of a large number of physicians and nurses working in the hospital. The data was commented upon by the directors and the members of the Ethical Committee of the hospital. The solutions suggested to meet the needs expressed by patients were published in the newsletter and the web site of the Institute, followed by articles on the Volunteer organizations, patients' right Associations newsletters and by popular magazines. This could be a first step in the direction of real change in cancer patient care.