American Journal of Respiratory and Critical Care Medicine

Rationale: Advances in the management of cystic fibrosis have led to a significant improvement in survival, although marked differences between individuals are still observed. The value of patient-reported health-related quality of life scores in predicting survival in adults with cystic fibrosis is unknown.

Objectives: To evaluate whether patient-reported health-related quality of life could predict survival in cystic fibrosis.

Methods: From 1996 to 1997 a consecutive series of 223 patients were recruited to evaluate the Cystic Fibrosis Quality of Life Questionnaire. Demographic (age, sex), clinical (FEV1% predicted, body mass index, diabetes, B. Cepacia complex, intravenous access device, nutritional and lung transplant status) and health-related quality of life variables were recorded (Cystic Fibrosis Quality of Life Questionnaire and the SF-36). These data were used as baseline measures to explore the prognostic association of health-related quality of life and subsequent survival.

Measurements and Main Results: At the census date (December 31, 2006) 154 (69.1%) adults were alive, 66 (29.6%) had died, and three (1.3%) were lost to follow-up. Cox proportional hazards models and bootstrapping procedures examined if health-related quality of life domains predicted survival after adjusting for the demographic and clinical factors. The physical functioning domain of the Cystic Fibrosis Quality of Life Questionnaire and the pain domain of the Short Form-36 had the strongest statistical associations with survival.

Conclusions: Aspects of patient-reported quality of life serve as prognostic measures of survival beyond a number of previously known factors in cystic fibrosis. This needs to be investigated further in a larger longitudinal study.

Scientific Knowledge on this Subject

Advances in the care of patients with cystic fibrosis (CF) have led to an improvement in survival. Even so, survival among people with CF varies substantially. Whether domains of health-related quality of life (HRQoL) can predict survival is unknown.

What This Study Adds to the Field

Aspects of patient-reported HRQoL serve as prognostic measures of survival beyond a number of previously known factors in cystic fibrosis.

Cystic fibrosis (CF) is the most common recessively inherited condition in the white population (1). It is a life-threatening disease primarily involving the respiratory and gastrointestinal systems. Advances in clinical care have led to a significant improvement in survival (2). Even so, survival among people with CF varies substantially. Commonly found risk factors for survival include birth cohort, poor pulmonary function, and the presence of specific airway pathogens (210). The identification of additional, modifiable prognostic factors may have important implications for guiding the management of patients with CF in clinical practice or for the design of intervention research in this disease group.

Subjective health reports and domains of patient-reported health-related quality of life (HRQoL) are important predictors of survival in healthy populations (11, 12). The reasons for this are not fully understood. Over recent years, accumulating reports have suggested that there is an association between HRQoL and survival in a range of disease conditions, particularly in cancer and cardiovascular disease (1322). However, not all studies have reported an effect on mortality (23, 24) and many have suffered from inadequate adjustment, failing to account for known clinical predictors of survival required to demonstrate an independent survival effect of HRQoL (13). Therefore, the prognostic value of patient-reported measures remains uncertain, and any predictive power of HRQoL may vary according to the nature of the disease. Whether domains of patient-reported HRQoL are able to predict survival in cystic fibrosis is unknown. This work was an exploratory study to evaluate whether patient-reported domains of the Cystic Fibrosis Quality of Life Questionnaire (25) or the Short Form-36 (SF-36) (26) could predict survival in cystic fibrosis after controlling for key biomedical variables. Some of the results of this study have been previously reported in the form of an abstract (27).

Subjects and Procedures

Consecutive patients who attended one of two Adult Cystic Fibrosis Units in the United Kingdom between September 1996 and December 1997 were approached to take part in a study evaluating the validity of the Cystic Fibrosis Quality of Life (CFQoL) instrument. Demographic, clinical, and HRQoL variables were assessed during the visit. These data were used as baseline measures to explore the prognostic association of HRQoL (CFQoL and SF-36) and subsequent survival. Patients gave consent to continuing access to their clinical records and these have been investigated to identify their vital status at the end of 2006.

The vital status of patients was identified from clinical records held at the CF Units. Time to death was measured from the time the person entered the study (and completed baseline HRQoL questionnaires) until date of death or 31 December 2006. Patients were classified as alive if they had attended an appointment, or had contact with a member of the medical team on or after this date. The data for patients whose status was unknown on 31 December 2006 were censored on the date of their last attendance at the CF Unit.


At entry to the original study, subjects' demographic, clinical, and HRQoL variables were collected. Age, sex, FEV1% predicted, body mass index, whether the person had diabetes, B. Cepacia complex, or an intravenous access device were recorded. Nutritional status (no oral calorie supplements, prescribed oral calorie supplements, or prescribed enteral tube feeds) and lung transplant status (not on waiting list, on waiting list, or post-transplant) were also documented.

Quality of Life was measured using the Cystic Fibrosis Quality of Life Questionnaire (CFQoL) and the Short Form 36 (SF-36). The CFQoL measures nine domains of functioning: physical functioning, social functioning, emotional responses, treatment issues, chest symptoms, body image, interpersonal relationships, career issues, and future concerns. The psychometric properties of the instrument were evaluated on this dataset. Internal reliability of the domains was demonstrated using Cronbach's α coefficients (range, 0.72–0.92; median, 0.89) and item to total domain score correlations. Concurrent validity with three appropriate SF-36 domains (range, r = 0.64–0.74), known groups validity between different levels of disease severity, sensitivity across transient changes in health (effect size range, d = 0.56–1.95) and test-retest reliability (r = 0.74–0.96; median, 0.91) were found to be robust. Each domain has a minimum score of 0 and a maximum score of 100, with higher scores reflecting a better quality of life (25).

The SF-36 comprises nine domains: physical functioning, role limitations physical, social functioning, mental health, role limitations emotional, bodily pain, energy and vitality, general health perceptions and changes in health. The instrument was evaluated for use with a United Kingdom CF population in this dataset. The domain structure was robust: internal reliability of the domains was demonstrated using Cronbach's α coefficients (range 0.82–0.91, median 0.84) and item to total domain score correlations were strong. Known groups validity between different levels of disease severity was noted for some domains. Each domain has a minimum score of 0 and a maximum score of 100, with higher scores reflecting a better quality of life (26).

Statistical Methods

Analyses were performed in Intercooled STATA 9. Summary statistics were obtained and then survival analysis was undertaken using Cox Proportional Hazards regression, including a check of the assumption of proportional hazards. Where there were concerns about this assumption, stratification was used to improve the model. The block of demographic and clinical variables was forced into all models. These demographic and clinical variables (excepting access device, which is indicative of more severe disease) have been shown to predict survival in CF epidemiological studies (210) and all 9 variables have been shown to explain domains on the CFQoL (28).The additional prognostic value of each of the 18 HRQoL variables was then investigated one at a time.

To investigate the relative importance of the HRQoL domains, an exploratory analysis was performed using backward elimination on all the 18 HRQoL variables, while retaining the block of demographic and clinical variables. The robustness of this analysis was checked in several ways. First, given the problems with automatic model selection and potential overfitting, bootstrapping in R1.9.0 was undertaken (2931). Bootstrapping is a method of resampling, with replacement, many times from the observed data (i.e., some cases appear more than once in a bootstrap sample whereas others are not present). This allows the performance of statistical methods to be checked by applying them to a large number of samples. The backward elimination was run on 50 bootstrap replications of samples of n = 223. If a statistically significant coefficient is due to a genuine effect, one would expect it to be statistically significant in a high proportion of the 50 replicated samples. Although this method cannot overcome all the problems caused by small samples it does provide additional information about the generalizability of the results, reducing the likelihood of making spurious conclusions based on models that may not be stable (e.g., due to sample size, automatic model selection or type I errors from multiple testing). Second, since B. cepacia complex is known to be a strong predictor of survival, the multivariable model was refitted, omitting the 25 patients who were colonized with this pathogen at baseline. Third, a stepwise algorithm was used (allowing entry and removal of HRQoL variables at each step) with the block of 9 demographic and clinical variables forced in. Fourth, an analysis of efficient score residuals was used to check for influential observations.

A total of 223 patients took part in the original study and their baseline demographical and clinical characteristics are presented in Table 1. The HRQoL scale information is summarized in Table 2. There were moderate correlations (r ≥ 0.5) between some HRQoL domains but the patterns are complex to summarize succinctly. See Table E1 in the online supplement for the full correlation matrix, including some key patterns described here. Within the SF-36, physical functioning, role physical, social functioning, and energy tended to form an intercorrelated subgroup. Within the CFQoL, physical functioning, social functioning, treatment, chest symptoms, and emotional functioning also tended to be an intercorrelated subgroup. Furthermore, there were correlations between these two subgroups. There were no other clear patterns of association. In particular, pain, changes in health, body image, career issues, and future concerns had very low correlations with other domains.


Baseline Variables

Number of Deaths
Males102 (46)26
Females121 (54)40
Age, yr25.1 (7.1) [range 14–52]
FEV1, %55.0 (23.5) [range 12.3–118.4]
 Mild disease60 (27)5
 Moderate disease97 (43)21
 Severe disease66 (30)40
BMI, kg/m220.8 (2.5) [range 15.5–30.4]
IV access device fitted63 (28)28
No IV access device fitted160 (72)38
B. Cepacia complex25 (11)15
No B. Cepacia complex198 (89)51
Diabetes49 (22)24
No diabetes174 (78)42
Nutritional supplement
 None103 (46)25
 Oral supplement92 (41)26
 Enteral feeds28 (13)15
Transplant status
 Not on list196 (89)50
 Transplant list13 (5)10
14 (6)

Definition of abbreviations: BMI = body mass index; IV = intravenous.

Values are n (%) or mean (SD) [range] unless otherwise indicated.


HRQoL Scale

Mean (SD)

Floor, n (%)

Ceiling, n (%)
 Physical functioning83.2 (18.5)2 (0.9)39 (17.5)
 Social functioning84.6 (21.4)1 (0.4)88 (39.5)
 Treatment issues74.5 (24.3)2 (0.9)46 (20.6)
 Chest symptoms65.7 (26.9)5 (2.2)30 (13.5)
 Emotional responses79.7 (17.8)1 (0.4)21 (9.4)
 Body image65.1 (24.9)1 (0.4)30 (13.5)
 Relationships62.4 (21.8)1 (0.4)7 (3.1)
 Career issues59.6 (28.4)3 (1.3)27 (12.1)
 Future concerns44.9 (23.6)06 (2.7)
 Physical function76.3 (24.0)1 (0.4)45 (20.2)
 Role limitation physical72.9 (38.9)37 (16.6)134 (60.1)
 Social function80.4 (23.8)0102 (45.7)
 Mental health73.7 (18.1)09 (4.0)
 Role limitation mental77.0 (37.0)31 (13.9)151 (67.7)
 Pain83.2 (21.3)1 (0.4)104 (46.6)
 Energy and vitality58.4 (23.1)1 (0.4)6 (2.7)
 Changes in health55.0 (21.9)2 (0.9)22 (9.9)
 Health perceptions
43.4 (23.7)
2 (0.9)
2 (0.9)

Definition of abbreviations: CFQoL = cystic fibrosis quality of life; HRQoL = health-related quality of life; SF-36 = short form-36.

At the census date, 154 (69.1%) adults were known to be alive, 66 (29.6%) had died (all CF-related deaths), and 3 (1.3%) were lost to follow-up. Further information about the known deaths is presented in Table 1. Figure 1 shows an unadjusted Kaplan-Meier curve with censored observations identified. In initial analyses there was some evidence that the proportional hazards assumption was invalid for two variables, both of which were statistically significant. The first variable was FEV1% predicted: as time progressed, the effect on the hazard function of low baseline FEV1% predicted was more pronounced than would be expected. The second variable was B. cepacia complex status where the effect on the hazard function tended to increase over time. (The hazard ratios were 0.97 and 3.00 respectively, although these should be interpreted with caution because of the problems with the model assumptions.) A stratified analysis was therefore performed for all models in the following way: FEV1% predicted was recoded as disease severity: mild disease (FEV1 >70%), moderate disease (FEV1 = 40–70%), severe disease (FEV1 <40%), and combined with the presence or absence of B. cepacia complex, giving six strata. The proportional hazards assumption was reasonable for all subsequent models. It may be more clinically acceptable to include a fourth category of “normal” lung function (FEV1 ≥100%); however, there were only nine patients in this category and the hazard functions appeared similar for normal (FEV1% predicted ≥100) and mild cases (FEV1% predicted 70–99).

Table 3 presents the estimates of hazard ratios (HR). Examining one HRQoL variable at a time, there was a general trend that higher HRQoL scores were associated with better survival for all HRQoL domains except for the psychosocial domains of mental health, body image, interpersonal relationships, career issues, and future concerns. Statistically significant associations were obtained for physical functioning, chest symptoms, and emotional responses on the CFQoL, and for pain and changes in health on the SF-36. After Bonferroni correction (P < 0.0028), physical functioning, pain, and changes in health were still significant. However, the model from backward elimination retained only two of these domains, each with HR less than one. These were physical functioning on the CFQoL (HR = 0.96; 95% CI [confidence interval] 0.93 to 0.98; P < 0.001) and pain on the SF-36 (HR = 0.98; 95% CI 0.97 to 1.00; P = 0.019). The bootstrapping results are summarized in Table 3 (for further details, see Table E2). They confirm the importance of the physical functioning and pain domains, each of which was retained in at least 40 of 50 replications. The change in health domain was retained over half of the time. Unexpectedly, the backward elimination procedure also retained social functioning from the CFQoL, but with HR greater than one (HR = 1.03; 95% CI 1.01–1.04; P < 0.006). This was retained in the bootstrapping half of the time, always with HR greater than one. See Table E3 for coefficients of all variables in the multivariable model.



Bootstrapping Results
HRQoL Domain
HR (95% CI)
P Value
HR <1
HR >1
 Physical functioning0.97 (0.96 to 0.99)<0.001420
 Social functioning0.99 (0.98 to 1.00)0.239025
 Treatment issues0.99 (0.98 to 1.00)0.16571
 Chest symptoms0.98 (0.97 to 0.99)0.005120
 Emotional responses0.98 (0.97 to 1.00)0.03252
 Body image1.00 (0.99 to 1.01)0.916017
 Relationships1.00 (0.99 to 1.01)0.99204
 Career issues1.00 (0.98 to 1.01)0.40142
 Future concerns1.00 (0.98 to 1.01)0.665110
 Physical function0.99 (0.98 to 1.00)0.06945
 Role limitation physical0.99 (0.98 to 1.00)0.172018
 Social function0.99 (0.98 to 1.00)0.115112
 Mental health1.00 (0.98 to 1.01)0.601019
 Role limitation mental0.99 (0.98 to 1.00)0.10365
 Pain0.98 (0.97 to 0.99)<0.001400
 Energy and vitality0.99 (0.98 to 1.00)0.112211
 Changes in health over the last 12 mo0.98 (0.97 to 0.99)0.002270
 General health perceptions
0.99 (0.98 to 1.01)

Definition of abbreviations: CFQoL = cystic fibrosis quality of life; CI = confidence interval; HR = hazards ratio; HRQoL = health-related quality of life; SF-36 = Short Form-36.

*Model including one HRQoL variable at a time. All models include the block of nine demographic and clinical variables.

Number of times (from 50 bootstrap replications) that HRQoL domain was statistically significant (P < 0.05). Results distinguish between cases where HR is less than or greater than one.

The robustness checks revealed no obvious influential observations, and removal of the patients with B. cepacia complex made little difference to the findings. When using the stepwise procedure, the choice of entry of the first HRQoL variable was marginal because the P values for physical functioning and pain domains were very similar and very low. Using the automated procedure, pain was selected first, and the change in health domain was added: both variables were subsequently retained. If physical functioning was entered first, the resulting model was identical to that of the backward elimination procedure (i.e., it contained physical functioning, pain, and social functioning). This model was also selected when the robust variance option was used in the automated stepwise procedure.

This work provides persuasive evidence that aspects of patient-reported quality of life serve as prognostic measures of survival beyond a number of previously known factors in CF. Specifically, the physical functioning domain of the CFQoL and the pain domain of the SF-36 are important predictors of survival, even in the same multivariable model. There is suggestive evidence that changes in health may be important and also that social functioning (measured by the CFQoL) may add information to the survival model when the physical functioning domain is already present.

Because the HRQoL predictor variables tend to be correlated, care must be taken in interpreting the coefficients in a multivariate model. When domains were considered one at a time, the chest symptoms and emotional responses domains of the CFQoL also appeared to be potentially important predictors of survival. However, given the correlations between physical functioning, chest symptoms, and emotional responses, it is likely that physical functioning is the most important. This is consistent with the bootstrapping results. The potential importance of the changes in health domain suggests that longitudinal models, which take into account changes over time, may be better at predicting survival.

The insights that patients have concerning their health and how they report these are important. Even with severe disease, many people report a good quality of life across many HRQoL domains. However, when adults with CF rate their physical functioning, they appear to be aware of something important that is not identified by traditional risk factors. It is possible that patients' scores reflect a more accurate perception of their general health/disease severity or they may act as a marker for a yet undetected prognostic factor. Indeed, in studies that have evaluated specific HRQoL domains rather than total scores, patient-reported physical functioning has been shown to be a predictor of survival in other chronic conditions (1315). It is important to note that, whereas the physical functioning domain on the CF-specific scale (the CFQoL) was statistically significant, the association between the physical functioning domain of the SF-36 and survival was weaker and not statistically significant. This highlights one of the benefits of using disease-specific scales. Another advantage for these data is that ceiling effects, although common in HRQoL scales (32), tended to be less pronounced in the CFQoL sample than in the SF-36 sample. Nonetheless, there are substantial ceiling effects and it is difficult to predict how these may have affected the results. For example, it is plausible that 46% of the sample did not experience pain. For other domains it is less clear whether values at ceiling reflect true values for the underlying construct or whether they are an artifact of the domain. Where ceiling effects are an artifact of the scale it is possible that actual effects have been underestimated.

The results concerning social functioning are counterintuitive and must be interpreted with caution. The multivariable modeling suggests that, after physical functioning is taken into account, lower social functioning scores predict better survival. The results from bootstrapping are somewhat ambivalent: social functioning was retained in 50% of the replications but always with an HR greater than one. This apparent effect could therefore be a Type 1 error, an artifact of this data set, or the results of overfitting from model selection. However, if this is a real effect, its interpretation requires a detailed examination of the social functioning items on the CFQoL. This is because social functioning is defined in different ways on different HRQoL instruments (items for key domains are presented in Table 4). The CFQoL asks about a person's enjoyment of life through socializing, whether they are more cautious and/or avoid going out or visiting friends because of their CF. It is feasible that social isolation is protective if patients are vulnerable to infection when visiting friends or socializing with other patients with CF. Furthermore, patients who report low social functioning scores may also engage in more cautious lifestyle behaviors.


Physical functioning (CFQoL)
 I have had difficulty doing heavy physical jobs. For example; digging, moving furniture, washing the car, vacuuming etc.
 My CF has prevented me from getting out of the house to run errands. For example; paying bills, posting a letter, doing light shopping etc.
 Because of my CF, it has been difficult for me to do light tasks around the house. For example; preparing a light snack, washing up, picking up the mail etc.
 Getting around the house has been difficult, because of my CF.
 CF has made it difficult to move from my bed or my chair.
 Despite CF, I have got around and done what I like.
 There are places that I would like to have gone, but didn't because of my CF.
 My CF has limited the type of sports and exercise I have been able to do.
 My CF has made me feel lacking in energy.
 I have found that my physical functioning and mobility have affected my quality of life by making life less enjoyable.
Social functioning (CFQoL)
 I have avoided going out and socializing because of my CF
 Because of my CF I have tended to avoid visiting friends
 When I have been out socializing I have behaved more cautiously than I would like to because of my CF
 I find that the way in which CF affects my socializing interferes with my overall enjoyment of life
Pain (SF-36)
 How much bodily pain have you had?
 How much did pain interfere with your normal work (including work both outside the home and at home)?
Changes in health (SF-36)
 Compared to one year ago, how would you rate your health in general now?

Definition of abbreviations: CF = cystic fibrosis; CFQoL = cystic fibrosis quality of life; SF-36 = short form-36.

The result that pain can predict survival in adults with CF is intriguing. The literature concerning pain in CF is sparse and it is noteworthy that pain is not measured by any of the CF-specific HRQoL scales. The nature of the pain is unknown, as the two pain items on the SF-36 are general (asking about the presence of pain and limitations due to pain—see Table 4); even so, they were able to predict survival. Just over one-half the sample (54%) reported pain, although there was no clear association between the SF-36 pain score and any of the other HRQoL predictor variables. Interestingly, in a large-scale population-based study in the UK, people who reported widespread pain had an increased risk of death, mainly from cancer, over the subsequent eight years (33).

A few studies, predominantly with small samples and different methods of measurement, suggest that one-half to two-thirds of children and adults with CF report pain on a regular basis (3439), which is associated with a poorer quality of life (3841). In a large sample of adults with CF, one-third reported episodes of pain as strong to severe, whereas 11% reported severe pain (39). There are many aspects of CF that may cause pain, with headache, backache, abdominal, chest, and limb pain most frequently reported. The literature contains inconsistent reports as to whether pain is underreported by people with CF and underestimated by health professionals (34, 37, 42). However, these current data suggest that pain is worthy of further investigation and draw attention to the importance of routinely assessing pain in people with CF.

The mechanism/s underlying the association between patient-reported aspects of HRQoL and length of survival in CF are not clear, but deserve closer investigation. Since the collection of the baseline data for this study, several additional risk factors have been proposed for CF including the annual rate of pulmonary exacerbations (5), sex (5, 43), CFTR genotype (7, 44), CF-related diabetes (5, 45), stature (46), and household income (47). Future work should endeavor to study large patient cohorts, adjusting for all proposed predictors of mortality (including FEV1) in ways that are both clinically and statistically robust. These models should also include time-dependent variables.

The results of our study add to the growing realization of the importance of patient-reported outcomes in the management of cystic fibrosis. They may also influence the choice of outcomes in randomized controlled trials, improve stratification of patients and interpretation of results (48). Often, the ultimate aim of a therapy is that it will increase survival, therefore surrogate markers of survival (e.g., FEV1), which are clinically important and sensitive to the treatment, are used as endpoints. This work provides credibility for the use of patient-reported measures in clinical trials in CF, as it provides evidence that patient-reported outcomes can predict mortality.

1. Kerem BS, Rommens JM, Buchanan JA, Markiewicz D, Cox TK, Chakravarti A, Buchwald M, Tsui LC. Identification of the cystic fibrosis gene: genetic analysis. Science 1989;245:1073–1080.
2. Dodge JA, Lewis PA, Stanton M, Wilsher J. Cystic fibrosis mortality and survival in the UK 1947–2003. Eur Respir J 2007;29:522–526.
3. Dodge JA, Morison S, Lewis PA, Coles EC, Geddes D, Russell G, Littlewood JM, Scott MT. Incidence, population and survival of cystic fibrosis in the UK (1968–95). Arch Dis Child 1997;77:493–496.
4. Elborn JS, Britton JR, Shale DJ. Cystic fibrosis: current survival and population predictions until the year 2000. Thorax 1991;46:881–885.
5. Liou TG, Adler FR, Fitzsimmons SC, Cahill BC, Hibbs JR, Marshal BC. Predictive 5 year survivorship model of cystic fibrosis. Am J Epidemiol 2001;153:345–352.
6. Courtney JM, Bradley J, McCaughan J, O'Connor TM, Shortt C, Bredin CP, Bradbury I, Elborn JS. Predictors of mortality in adults with cystic fibrosis. Pediatr Pulmonol 2007;42:525–532.
7. Lai HJ, Cheng Y, Cho H, Kosorok MR, Farrell PM. Association between initial disease presentation, lung disease outcomes, and survival in patients with cystic fibrosis. Am J Epidemiol 2004;159:537–546.
8. Corey M, Farewell V. Determinants of mortality from cystic fibrosis in Canada, 1970–1989. Am J Epidemiol 1996;143:1007–1017.
9. Kalish LA, Waltz DA, Dovey M, Potter-Bynoe G, McAdam AJ, LiPuma JJ, Gerard C, Goldmann D. Impact of Burkholderia dolosa on lung function and survival in cystic fibrosis. Am J Respir Crit Care Med 2006;173:421–425.
10. Gibson RL, Burns JL, Ramsey BW. Pathophysiology and management of pulmonary infections in cystic fibrosis. Am J Respir Crit Care Med 2003;168:918–951.
11. Idler EL, Benyamini Y. Self-rated health and mortality: a review of 27 community studies. J Health Soc Behav 1997;38:21–37.
12. Benyamini Y, Idler EL. Community studies reporting association between self-rated health and mortality: additional studies, 1995–1998. Res Aging 1999;21:392–401.
13. Fielding R, Wong WS. Quality of life as a predictor of cancer survival among Chinese liver and lung patients. Eur J Cancer 2007;43:1723–1730.
14. Coates A, Porzsolt F, Osoba D. Quality of life in oncology practice; prognostic value of EORTC QLQ-C30 scores in patients with advanced malignancy. Eur J Cancer 1997;33:1025–1030.
15. Fang FM, Tsai WL, Chiu HC, Kuo WR, Hsiung CY. Quality of life as a survival predictor for espohageal squalamous cell carcinoma treated with radiotherapy. Int J Radiat Oncol Biol Phys 2004;58:1394–1404.
16. Efficace F, Bottomley A, Coens C, Van Steen K, Conroy T. Schöffski P, Schmoll H, Van Cutsem E, Köhne CH. Does a patient's self-reported health-related quality of life predict survival beyond key biomedical data in advanced colorectal cancer? Eur J Cancer 2006;42:42–49.
17. Montazeri A, Milroy R, Hole D, McEwen J, Gillie CR. Quality of life in lung cancer patients as an important prognostic factor. Lung Cancer 2001;31:233–240.
18. Maisey NR, Norman A, Watson M, Allen MJ, Hill ME, Cunningham D. Baseline quality of life predicts survival in patients with advanced colorectal cancer. Eur J Cancer 2002;38:1351–1357.
19. Koch CG, Li L, Lauer M, Sabik J, Starr NJ, Blackstone EH. Effect of functional health-related quality of life on long-term survival after cardiac surgery. Circulation 2007;115:692–699.
20. Bardage C, Isacson D, Pedersen NL. Self-rated health as a predictor of mortality among persons with cardiovascular disease in Sweden. Scand J Public Health 2001;29:13–22.
21. Westin L, Nilstun T, Carlsson R, Erhardt L. Patients with ischemic heart disease: quality of life predicts long-term mortality. Scand Cardiovasc J 2005;39:50–54.
22. Lenzen MJ, Reimer WJM, Pedersen SS, Boersma E, Maier W, Widimsky P, Simoons ML, Mercado NF, Wijns W. The additional value of patient-reported health status in predicting 1 year mortality after invasive coronary procedures: a report from the Euro Heart Survey on Coronary Revascularisation. Heart 2007;93:339–344.
23. Ringdal GI, Götestam KG, Kaasa S, Kvinnsland S, Ringdal K. Prognostic factors and survival in a heterogeneous sample of cancer patients. Br J Cancer 1996;73:1594–1599.
24. Herndon JE, Fleishman S, Kornblith AB, Kosty M, Green MR, Holland J. Is quality of life predictive of the survival of patients with advanced nonsmall cell lung carcinoma? Cancer 1999;84:333–340.
25. Gee L, Abbott J, Conway SP, Etherington C, Webb AK. Development of a disease specific health related quality of life measure for adults and adolescents with cystic fibrosis. Thorax 2000;55:946–954.
26. Gee L, Abbott J, Conway SP, Etherington C, Webb AK. Validation of the SF-36 for the assessment of quality of life in adolescents and adults with cystic fibrosis. J Cyst Fibros 2002;1:137–145.
27. Abbott MA, Hart A, Conway SP, Webb AK. Can patient-reported health-related quality of life predict survival in cystic fibrosis? J Cyst Fibros 2007;6:S77.
28. Gee L, Abbott J, Hart A, Conway SP, Etherington C, Webb AK. Associations between clinical variables and quality of life in adults with cystic fibrosis. J Cyst Fibros 2005;4:59–66.
29. Babyak MA. What you see may not be what you get: a brief, nontechnical introduction to overfitting in regression-type models. Psychosom Med 2004;66:411–421.
30. Davison AC, Hinkley DV. Bootstrap methods and their applications. Cambridge: Cambridge University Press, 1997.
31. Harrell FE. Regression modeling strategies: with applications to linear models, logistic regression and survival analysis. New York: Springer, 2001.
32. Quittner A, Buu A, Messer M, Modi A, Watrous M. Development and validation of the cystic fibrosis questionnaire in the United States: a health-related quality-of-life measure for cystic fibrosis. Chest 2005;128:2347–2354.
33. McFarlane G, Mcbeth J, Silman A. Widespread body pain and mortality: prospective population based study. BMJ 2001;323:1–5.
34. Epker J, Maddrey AM, Rosenblatt R. Pain and pain-related impairment in adults with cystic fibrosis. J Clin Psychol Med Settings 1999;6:393–403.
35. Hubbard PA, Broome ME, Antia LA. Pain, coping and disability in adolescents and young adults with cystic fibrosis: a web-based study. Pediatr Nurs 2005;31:82–86.
36. Koh JL, Harrison D, Palermo TM, Turner H, McGraw T. Assessment of acute and chronic pain symptoms in children with cystic fibrosis. Pediatr Pulmonol 2005;40:330–335.
37. Ravilly S, Robinson W, Suresh S, Wohl ME, Berde CB. Chronic pain in cystic fibrosis. Pediatrics 1996;98:741–747.
38. Palermo TM, Harrison D, Koh JL. Effect of disease-related pain on the health-related quality of life of children and adolescents with cystic fibrosis. Clin J Pain 2006;22:532–537.
39. Festini F, Ballarin S, Codamo T, Doro R, Loganes C. Prevalence of pain in adults with cystic fibrosis. J Cyst Fibros 2004;3:51–57.
40. Chastain D. Chronic pain in CF: associated beliefs and behaviours. Pediatr Pulmonol 2000;30:116–117.
41. Bilton D, Landy A, Gunn E, Saunders M. Is pain and discomfort common in CF adults? J Cyst Fibros 2001;1:320.
42. Eksterowicz N. Pain management in cystic fibrosis. Pediatr Pulmonol 2000;30:114–115.
43. Rosenfield M, Davis R, Fitzsimmons S, Pepe M, Ramsey B. Gender gap in cystic fibrosis mortality. Am J Epidemiol 1997;145:794–803.
44. McKone EF, Goss CH, Aitken ML. CFTR genotype as a predictor of prognosis in cystic fibrosis. Chest 2006;130:1441–1447.
45. Bismuth E, Laborde K, Taupin P, Velho G, Ribault, Jennane F, Grasset E, Sermet I, De Blic J, Lenoir G, Robert J-J. Glucose tolerance and insulin secretion, morbidity, and death in patients with cystic fibrosis. J Pediatr 2008;152:540–545.
46. Beker LT, Russek-Cohen E, Fink RJ. Stature as a prognostic factor in cystic fibrosis survival. J Am Diet Assoc 2001;101:438–442.
47. O'Connor GT, Quinton HB, Kneeland T, Kahn R, Lever T, Maddock J, Robichaud P, Detzer M, Swartz DR. Median household income and mortality rate in cystic fibrosis. Pediatrics 2003;111:e333–e339.
48. Federal Drugs Agency. Guidance for industry. Patient-reported outcome measures: Use in medical product development to support labeling claims. Draft guidance, February 2006. [Accessed Oct 10, 2007.] Available at:
Correspondence and requests for reprints should be addressed to Janice Abbott, Ph.D., C.Psychol., Faculty of Health, University of Central Lancashire, Corporation Street, Preston PR1 2HE, United Kingdom. E-mail:


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record