The concept of quality of life related to health (HRQoL) is defined with regard to the way in which illness (as a source of pain, physical dysfunction and discomfort) imposes limitations or alterations on everyday behaviour, social activities and psychological wellbeing, as well as in other aspects of personal daily life(Reference Hendry and McVittie1).
The measurement of quality of life brings a holistic dimension to the burden of a clinical state or to the response to an operation. The relationship between quality of life and nutritional status is not well studied. Furthermore, measuring HRQoL is a complex process, being, as it is, a subjective, multifactor construct responsive to individual expectations in different facets of life. The way in which HRQoL is affected by the nutritional status of the patient is a subject of constant interest and permanent debate. It is all too well known that an impoverishment of nutritional status leads to a decrease in physiological function, increasing the risk of complications and septic death(Reference Rantanen, Harris and Leveille2, Reference Cabioglu, Bilgic and Deniz3), that there is a significant correlation between nutrition and alterations in muscular, immune and cognitive functions(Reference Pearson, Schlettwein-Gsell and Brzozowska4) and therefore that an improvement in nutritional status is an influencing factor in the improvement of physiological function(Reference Bourdel-Marchasson, Joseph and Dehail5, Reference Magri, Borza and del Vecchio6).
The necessity and importance of the measurement of HRQoL, both general and specific, tied to a definite concept, can be justified on the basis of studies which show that perceived health is independently associated with medium-term mortality(Reference O'Reilly, Rosato and Patterson7, Reference Fillenbaum, Burchett and Kuchibhatla8). These specific instruments, designed to relate a patient's HRQoL to a specific pathology, have grown in importance in recent years. They also provide a subset of relevant data which point to a positive causality(Reference August9).
Consequently, the purpose and objective of the present study is to bring together those studies that relate HRQoL with nutritional status and examine the tools (questionnaires) that they use to investigate this relationship.
Methods
Bibliographic search
Given the hierarchical structure of medical subject heading (MeSH) terms, the terms ‘quality of life’, ‘nutritional status’ and ‘questionnaires’ were chosen and used in conjunction with the Boolean link ‘AND’.
The search was carried out from the earliest date possible (according to each database) until July 2007, the latest date considered in the present study.
In the only databases that permitted it, MEDLINE and EMBASE, the major (Majr) topic terms were used. These represent the most important concepts of an article and help to eliminate less relevant studies from the results, thereby increasing the sensitivity of the search (‘quality of life’ [Majr] AND ‘nutritional status’[Majr] AND ‘questionnaires’ [MeSH]).
‘Humans’ was used in all databases as a search limit.
Additionally, as a secondary search, the bibliographies of the selected articles were reviewed in order to identify studies not found by the primary search.
The databases MEDLINE (via PubMed), EMBASE, The Cochrane Library, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Institute for Scientific Information (ISI) Web of Science, Latin American and Caribbean Health Sciences Literature (LILACS) and Spanish Health Sciences Bibliographic Index (IBECS) were consulted.
Selection of the articles
The articles were selected via inclusion and exclusion criteria previously defined in a written protocol(Reference Wanden-Berghe, Sanz-Valero and Juan-Quilis10, Reference Wanden-Berghe, Sanz-Valero and Juan-Quilis11).
Inclusion criteria were:
(1) Use of at least one questionnaire that evaluates quality of life;
(2) Nutritional status of the studied individuals is, by whatever means, taken into account;
(3) Original articles published in peer-reviewed journals.
Excluded were studies that measured HRQoL using only clinical indicators.
Validity check
The studies, with no indication of the authors, journal or database origin, were checked for relevance by the three experts in nutrition (C. W.-B., J. M. Culebras and J. Alvarez) using a yes/no checklist(11).
Concordance analysis between the experts in nutrition (gold standard) and the obtained results gave the following results: observed agreement 90·00 (95 % CI 80·70, 99·30) %; κ 75 (95 % CI 52, 98) %; significance test 4·74 (P < 0·001); sensitivity 93·10 (95 % CI 83·88, 100) %; specificity 81·82 (95 % CI 59·03, 100) %. The silent percentage (relevant articles not found) and the noisy percentage (non-relevant articles found) were 5 (95 % CI 0, 11·75) % in both cases.
Special characteristics of the study
Although it is preferable to base a systematic review on prospective studies or studies with adequate follow-up periods, it was decided to include cross-sectional studies or studies with short follow-up periods if HRQoL had been studied using a valid questionnaire and the nutritional status of the observed patients had been taken into account. This limitation will be discussed later.
Results
Twenty-nine papers from MEDLINE, twenty-one from EMBASE, six from the Cochrane Library and thirteen from CINAHL were obtained. All the papers found in the bibliographic database ISI Web of Science had been previously collected. No articles were found in the databases LILACS or IBECS. After eliminating redundant papers, forty documents were obtained.
Agreement between the scientific documentation experts (J. S.-V., V. Juan-Quilis and R Ballester Añon; applying the most sensitive search formula) and the experts in nutrition reduced the number of studies to thirty-one(11).
A further study was discarded for measuring user satisfaction with nutrition services, rather than quality of life, and for not using a questionnaire that evaluated quality of life. Finally, twenty-eight documents on quality of life related to nutritional status were accepted(Reference O'Keefe, Emerling and Koritsky12–Reference McLeod, Taylor and O'Connor39) (Table 1).
Summary of the twenty-eight reviewed studies in chronological order according to year of publication
QoL, quality of life; M/W, men and women; EORTC QLQ-C-30, European Organisation for Research and Treatment of Cancer Quality of Life; EORTC QLQ-H&N35, European Organisation for Research and Treatment of Cancer Quality of Life – Head and Neck; SF-36, Short Form-36 Health Survey; HRQoL, health-related quality of life; WHOQOL-BREF; WHO Quality of Life-BREF; QoL-OS, Quality of Life focused on symptoms of oxidative stress; EuroQoL-5D, Euro Quality of Life 5 Dimensions; MNA Mini Nutritional Assessment; OHIP-EDENT, Oral Health Impact Profile – EDENT; PEG, percutaneous endoscopic gastrostomy; SCREEN, Seniors in the Community: Risk Evaluation for Eating and Nutrition; POQOLS, Paediatric Quality of Life Scale; PG-SGA, Patient-Generated Subjective Global Assessment; VAS, visual analogue scale; FAACT, Functional Assessment of Anorexia/Cachexia Therapy; BACRI, Bristol-Myers Anorexia Cachexia Recovery Instrument; COOP-WONCA, Dartmouth Primary Care Cooperative Information Project World Organization of National Colleges, Academies, and Academic Associations of General Practitioners/Family Physicians; FLIC, Functional Living Index-Cancer; LASA, linear analogue scale assessment.
It is worth noting how recent the studies are; the average age is 3·85 (95 % CI 2·62, 5·02) years, and the average obsolescence gives a value of 3 years and a Price index of 75 % (percentage of articles 5 years old or less).
The designs of the studied articles were: eight (28·57 %) clinical trials; eleven (39·29 %) prospective; seven (25·00 %) cross-sectional; two (7·14 %) retrospective. The disparity in design can be seen in the wide diversity of follow-up periods. The number of patients also varied widely, from a minimum of twelve to a maximum of 367.
Quality of life related to health and nutritional status
Although there are a considerable number of published studies on HRQoL, those that truly evaluate quality of life related to nutritional status are scarce. However, it is worth noting that of the articles relating HRQoL to nutritional status, eleven (39·27 %) had cancer as a pathological base(Reference Oates, Clark and Read13, Reference Murawa, Murawa and Oszkinis16, Reference Izutsu, Tsutsumi and Islam17, Reference Ravasco, Monteiro-Grillo and Vidal21, Reference Kennedy, Tucker and Ladas27, Reference Isenring, Bauer and Capra29, Reference Tomíska, Tomisková and Salajka30, Reference Ribaudo, Cella and Hahn33, Reference Van Bokhorst-de Van der Schuer, Langendoen and Vondeling35, Reference Bruera, Ernst and Hagen37, Reference Beller, Tattersall and Lumley38).
The review found no specific questionnaire that determined a direct link between HRQoL and nutritional status. However, three papers (10·71 %) detailed a significant correlation between nutritional status and HRQoL using a valid method for measuring quality of life(Reference Eriksson, Dey and Hessler19, Reference Keller23, Reference Isenring, Bauer and Capra29). Another article (3·57 %) referred to a possible relationship between HRQoL and nutrition(Reference Hickson and Frost25), but drew attention to other important factors, such as the risk of depression. A different study (3·57 %) mentioned how the ingestion of foodstuffs affects HRQoL(Reference Trabal, Leyes and Forga15), although a further paper(Reference Ribaudo, Cella and Hahn33) (3·57 %) found no significant effect between the results obtained using The Short Form-36 Health Survey (SF-36) questionnaire and nutritional intervention. In another, an association between a deteriorating HRQoL and severe malnutrition was seen(Reference Laws, Tapsell and Kelly34).
Among the reviewed papers, ten (35·71 %) found no type of relationship between nutritional status, or any type of nutrition, and HRQoL(Reference O'Keefe, Emerling and Koritsky12, Reference Murawa, Murawa and Oszkinis16–Reference Gramignano, Lusso and Madeddu18, Reference Allen20, Reference Scott, Beech and Smedley22, Reference Gollub and Weddle24, Reference Kennedy, Tucker and Ladas27, Reference Tidermark31, Reference Ribaudo, Cella and Hahn33).
Nine articles (32·14 %) recommended, or considered necessary, future prospective studies in order to completely clarify the correlation between HRQoL and nutritional status(Reference Kalaitzakis, Simrén and Olsson14, Reference Trabal, Leyes and Forga15, Reference Eriksson, Dey and Hessler19, Reference Allen20, Reference Keller23, Reference Johansen, Kondrup and Plum26, Reference Kennedy, Tucker and Ladas27, Reference Ribaudo, Cella and Hahn33, Reference Laws, Tapsell and Kelly34).
It is important to emphasise the study of Ravasco et al. (Reference Ravasco, Monteiro-Grillo and Vidal21), where the existence of a linear association (P < 0·05) between an increase in HRQoL and an improvement in nutritional status was demonstrated. The research of Isenring et al. (Reference Isenring, Bauer and Capra29) determined that 26 % (P < 0·001) of the appreciated variation in HRQoL is explained by changes observed in nutritional status measured with the ‘Patient-Generated Subjective Global Assessment’ (PG-SGA). By means of multivariate analysis Keller(Reference Keller23) showed that the association between nutritional risk and HRQoL is consistent, explaining the 44 % variation Hickson & Frost(Reference Hickson and Frost25) describe, concluding with the necessity for a tool that shows high sensibility to alterations of HRQoL and their relationship with nutritional status.
Questionnaire description and use
The questionnaires that were used in more than one article are: the European Organisation for Research and Treatment of Cancer Quality of Life questionnaire (EORTC-QLQ-C-30) and SF-36 on six occasions each; the Euro Quality of Life 5 Dimensions (EuroQoL-5D) on three occasions; linear analogue scale assessment (LASA) or visual analogue scales (VAS) on three occasions. In two studies, non-validated instruments were used to evaluate quality of life. The rest of the questionnaires were only used once.
It was observed that in one article (3·57 %) six different questionnaires were used to measure quality of life(Reference McLeod, Taylor and O'Connor39), in another (3·57 %) three questionnaires(Reference Tidermark31), in five articles (17·86 %) two were used(Reference Gramignano, Lusso and Madeddu18, Reference Scott, Beech and Smedley22, Reference Tomíska, Tomisková and Salajka30, Reference Van Bokhorst-de Van der Schuer, Langendoen and Vondeling35, Reference Bruera, Ernst and Hagen37) and in the rest only one.
Most of the questionnaires described in the studies measured quality of life in a generic way (SF-36; EuroQoL-5D; Dartmouth Primary Care Cooperative Information Project-World Organization of National Colleges, Academies, and Academic Associations of General Practitioners/Family Physicians (COOP-WONCA); LASA or VAS; Nottingham Health Profile; Physician Global Assessment (MD global); Quality of Well Being Scale; Sickness Impact Profile; Time Trade-off Technique; Vailas; WHO Quality of Life-BREF (WHOQOL-BREF)). Several were specific for cancer (EORTC QLQ-C-30; EORTC Head and Neck questionnaire (QLQ-H&N35); Functional Living Index-Cancer; Kurihara; Quality of Life focused on symptoms of oxidative stress) or for gastrointestinal pathology (Subjective Assessment of Quality of Life (Transplantation); Troidl; Visick scale).
Only three questionnaires that can be related to quality of life could be retrieved, two of them specific for anorexia and cachexia (Bristol-Myers Anorexia Cachexia Recovery Instrument; Functional Assessment of Anorexia/Cachexia Therapy) and one specific for patients with permanent home parenteral nutrition (Direct Questioning of Objectives) (Table 2).
Questionnaires used in reviewed articles
Discussion
In the documentary study the validity of the articles must be emphasised. The validity was confirmed both by the good results given by measuring the obsolescence, and by the excellent result of the Price index. The excellence and the current importance of the research articles are complementary, but nevertheless important factors, in those studies referring to the health sciences.
It must be underlined that the evaluation of HRQoL is circumscribed specifically within the scope of the investigation. Its use in common medical practice would help to obtain validated information about the impact of the illness or the treatment of the patient in daily life, both of which could be useful in decision making(Reference Lizán Tudela and Badia Llach40). Knowing HRQoL does not substitute the symptomatic, analytic and morphological evaluations, but complements them, introducing something as important as the patient's point of view about their perception of their own health(Reference Monés41). Quality of life assessment measuring the patient's experiences of the impact of disease and therapy, expectations and satisfaction should be the ‘gold standard’ as an independent end point in clinical trials(Reference Ravasco, Monteiro-Grillo and Vidal21, Reference Testa and Simonson42).
The undertaking of prospective studies of HRQoL in clinics improves the information about the patient, which, along with the diagnosis, provides important information about the patient's perception of the effect of treatment(Reference Oates, Clark and Read13, Reference Detmar, Muller and Schornagel43). Neither must it be forgotten that the objective is also to prioritise resources. Mathematical methods are applied to try to quantify the quality of life in relation to its usefulness (quality-adjusted life years) as a self-profit concept(Reference Zaragaza Monzón, Culebras and Gómez Candela44).
The review proves that HRQoL has been studied as a variable in the health-illness process(Reference Bowling45, Reference Milner46), and not only, as is frequently the case, as a covariable in pathological(Reference Kalaitzakis, Simrén and Olsson14, Reference Cabré Gelada and Gassull Duro47–Reference Eiser and Morse51) or surgical(Reference Ammerman, Watters and Clinch52–Reference Barreto Villela, Braghrolli Neto and Lima Curvello54) process studies, in pharmacological follow-ups(Reference Cuerda, Camblor and Bretón55), in relation to somatic(Reference Carlsson, Bosaeus and Nordgren56, Reference Huisman-de Waal, Schoonhoven and Jansenb57) or social(Reference Malone58, Reference Úbeda, Basagoiti and Alonso-Aperte59) aspects, or to support future recommendations(Reference Sawyer, Drew and Yeo60–Reference Baxter, Fayers and McKinlay62).
Limitations to the identified studies
The present review exposes the lack of homogeneity of the studies found, produced by certain limitations, namely the different questionnaires used, the diversity of pathologies, sample sizes, methodology and variation in follow-up, all of which do not permit meaningful meta-analysis, thus making direct comparison awkward, especially those studies that apply non-validated HRQoL questionnaires. Furthermore, two studies are retrospective in design and are susceptible to bias.
It is fundamental when designing these studies that possibly confusing variables are controlled, that interaction effects are recognised and that HRQoL is evaluated at different points in the evolution of the illness, the period in which patients are having the treatment or that these are matched up with a control group(Reference Wanden-Berghe63).
Limitations due to the questionnaires
No mention of the patient's acceptance of the HRQoL questionnaires used has been found in the studies reviewed. The complexity of these tools or their use can be the cause of disinterestedness, partial fulfilment or desertion on the part of the participants. This conformity is a crucial methodological requisite for avoiding skewed results(Reference Smith and Huntington64). It is possible that the structure and appearance of the questionnaires about HRQoL are considered less important than the final results, but if this circumstance is not properly managed, it will never be known if the results are influenced by the tool's design. On the other hand, it is convenient to limit the number of questionnaires used; some studies recommend not using more than three, if possible, or up to five in extreme cases(Reference Bottomley and Therasse65).
The use of validated and reliable measurements of HRQoL is essential. Ideally, any generic measurement of HRQoL should be replaced with a specific measurement that reflects the sensibility to the changes produced by the illness or by the influences related to the treatment. These questionnaires should not only have to be sensitive to the changes produced in the desired variable, but should also be acceptable to the patients(Reference Smith and Huntington64).
Quality of life and nutritional status
The relationship existing between nutritional status and HRQoL is becoming an important question not only in the study of oncological patients(Reference Marín Caro, Laviano and Pichard66, Reference García-Luna, Parejo Campos and Pereira Cunill67), but also in other pathologies(Reference de Miguel Díez, Izquierdo Alonso and Molina París68) and interventions(Reference Juan Samper, Ramón Capilla and Cantó Armengol69, Reference Galindo, Pérez de la Cruz and Cerezo70). The improvement of this correlation, as a consequence of an appropriate nutritional intervention, enables the reduction of the number of post-surgical complications(Reference Galindo, Pérez de la Cruz and Cerezo70, Reference Bozzetti, Braga and Gianotti71), shortens the recovery time and the length of hospital stay, improves tolerance to the treatment(Reference Fearon and Luff72, Reference Braga, Gianotti and Nespoli73) and even increases the rate of survival(Reference Marín Caro, Laviano and Pichard74–Reference Bozzetti, Cozzaglio and Biganzoli77), and with it a general decrease in morbidity(Reference Schneider, Veyres and Pivot78, Reference Farreras, Artigas and Cardona79).
On the other hand, as has been seen in the reviewed studies, the advice and nutritional follow-up given by professionals is related directly to the improvement in nutritional status, which will be related to the improvement in HRQoL(Reference Hickson and Frost25, Reference Ravasco, Monteiro-Grillo and Vidal80). It has been demonstrated, in head and neck neoplasm, that nutritional advice enables improvements in quality of life greater than those obtained by nutritional supplementation without advice(Reference Ravasco, Monteiro-Grillo and Marques Vidal81).
Now, the efficacy of nutritional advice as a positive influence on HRQoL depends on the possibility of adapting intervention to the specific need of each type of patient. Therefore, nutritional advice should be given by dedicated, specialised groups(Reference López Varela, Anido and Larrosa75, Reference Marín Caro, Laviano and Pichard82). Of special importance is the need for future studies that clarify the relationship between nutritional status and quality of life. This importance is recognised by studies included in the present review(Reference Kalaitzakis, Simrén and Olsson14, Reference Trabal, Leyes and Forga15, Reference Eriksson, Dey and Hessler19, Reference Allen20, Reference Keller23, Reference Johansen, Kondrup and Plum26, Reference Kennedy, Tucker and Ladas27, Reference Ribaudo, Cella and Hahn33, Reference Laws, Tapsell and Kelly34) and also in other publications that highlight the need to explore the relationship in greater detail.
Hence, the measurement of HRQoL with generic tools requires large sample sizes in order to demonstrate statistically significant differences and, in the majority of cases, these types of questionnaires are affected by uncontrolled external factors(Reference Testa and Simonson42, Reference Winkler83–Reference Genton, van Gemert and Pichard86). Ultimately, valid HRQoL measurement tools are dependent upon patient perception, the impact of the illness, the treatment, expectations and wellbeing. There should be an independent gold standard for all research projects and everyday medical practice.
A specific tool is needed: one that is sensitive to the measurement of HRQoL and can be self-administered quickly and easily on a regular basis. Nevertheless, it must be recognised that the development of a tool to detect, evaluate and monitor the influence of the pathological base is not an easy task.
Conclusion
Only three studies(Reference Eriksson, Dey and Hessler19, Reference Keller23, Reference Isenring, Bauer and Capra29) selectively focused on the relationship between nutritional status and quality of life, this evaluation being performed not by means of specific questionnaires but by statistical analysis of data obtained via validated questionnaires.
Acknowledgements
All the authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
This present investigation has been made possible thanks to a Nutricia grant.
All the authors contributed in the study, according to the Vancouver rules. Contributions of the authors were as follows: study concept and design, C. W.-B., J. S.-V. and V. E.-A.; bibliographical previous search, J. S.-V. and C. W.-B.; analysis and interpretation of data, analysis of nutritional contents, C. W.-B.; analysis of documental contents, J. S.-V; drafting of the manuscript, J. S.-V., I. C.-B., R. G.-W.-B.; elaboration of tables and databases, I. C.-B., R. G.-W.-B.; critical revision of the manuscript for important intellectual content, C. W.-B., V. E.-A.; statistical expertise, J. S.-V.; obtained funding, C. W.-B.; administrative, technical or material support, I. C.-B., R. G.-W.-B.; study supervision, C. W.-B., J. S.-V. and V. E.-A.
None of the authors has financial or other conflicts of interest.
