0
Clinical Investigations in Critical Care |

2-Month Mortality and Functional Status of Critically Ill Adult Patients Receiving Prolonged Mechanical Ventilation* FREE TO VIEW

Quality of Life After Mechanical Ventilation in the Aged Study Investigators
Author and Funding Information

Affiliations: *From the University of Pittsburgh School of Medicine, Pittsburgh, PA. ,  A list of study investigators is given in the Appendix.

Correspondence to: Lakshmipathi Chelluri MD, MPH, FCCP, University of Pittsburgh School of Medicine, 200 Lothrop St, 608 Scaife Hall, Pittsburgh, PA 15213; e-mail: chelluril@anes.upmc.edu


Affiliations: *From the University of Pittsburgh School of Medicine, Pittsburgh, PA. ,  A list of study investigators is given in the Appendix.


Chest. 2002;121(2):549-558. doi:10.1378/chest.121.2.549
Text Size: A A A
Published online

Study objectives: To describe the 2-month mortality and functional status of adult patients receiving prolonged (at least 48 h) mechanical ventilation (MV), and to identify patient characteristics that are associated with 2-month mortality.

Design: Prospective cohort study.

Setting: Four ICUs at a tertiary-care institution.

Patients: Eight hundred seventeen patients who received prolonged MV.

Interventions: None.

Measurements and results: Median age, sex distribution, and median Charlson comorbidity score of the 817 patients were 65 years, 45.8% women, and 1, respectively. The median scores on Katz Activities of Daily Living, Instrumental Activities of Daily Living Deficits, and Medical Outcomes Study Short-Form 36 surveys before hospitalization were 0, 1, and 50, respectively. Median APACHE (acute physiology and chronic health evaluation) III score and probability of hospital death for the cohort were 64 and 0.31, respectively. Median duration of MV was 9 days. Two-month mortality was 43%. Independent predictors of mortality at 2 months were age, comorbidities, and prehospital functional status. The adjusted odds of dying within 2 months increased 34% for each decade increase in age. Functional status deteriorated at 2 months compared to functional status prior to hospitalization, and 35% of the survivors were at risk for clinical depression. Among the 2-month survivors for whom the need for a caregiver was assessed, 78% had a caregiver.

Conclusions: Older age, in addition to functional status and comorbidities, was associated with increased mortality at 2 months. Functional status of survivors declined at 2 months.

Figures in this Article

Life expectancy has improved significantly in the past few decades because of advances in public health and medical care. The fastest growing segment of the population is elderly individuals at least 65 years old.1 Although the elderly constitute approximately 13% of the population, they use 33% of health-care resources.1Mechanical ventilation (MV) is a commonly used mode of support in the ICU.2The mean age of patients in the ICU and those receiving MV has also been increasing in the past few decades.3MV is an expensive modality, and patients requiring MV use more resources and have poorer outcomes compared to those patients who do not need mechanical ventilatory support.4Ely et al5reported similar hospital mortality for patients receiving MV and aged ≥ 75 years and patients < 75 years old. Other studies610 evaluating mortality after MV have shown a higher in-hospital mortality rate in older, compared to younger, patients receiving MV. However, the association of age with this in-hospital mortality, independent of other risk factors such as comorbidities and severity of illness, is not clear. In addition, sex may have an impact on mortality. Kollef et al11reported increased in-hospital mortality in women requiring MV, although severity of illness was a major independent predictor of mortality. Although many investigators evaluated the in-hospital outcomes of patients receiving MV, few report on post-hospital-discharge outcomes such as functional status, quality of life, and mortality. We previously documented that increased mortality from critical illness continues following hospital discharge, and measures of functional status of survivors deteriorate initially, but improve over a longer period.12 That study was limited, however, by focusing only on elderly critically ill patients and it did not include many patients receiving MV. As MV is burdensome and expensive, it would be helpful to have a better understanding of factors related to survival and quality of life after discharge from the hospital in patients receiving MV. Such knowledge could help physicians and patients in making decisions about life-sustaining therapy.

We therefore performed a prospective study of patients receiving MV for at least 48 h to study the association of age and sex on mortality and functional status at 2 months after the need for MV. We also describe the need for support after discharge from the hospital.

Study Site and Patient Selection

The Quality of Life After Mechanical Ventilation in the Aged (QOL-MV) study was conducted at the University of Pittsburgh Medical Center (UPMC) Health System, a tertiary-care facility. QOL-MV subjects were recruited from the medical, neurologic, trauma, and general surgical ICUs at the UPMC. QOL-MV is a study evaluating long-term (1 year) mortality and functional status of patients receiving prolonged MV. In this article, we report outcome at 2 months.

During a recruitment period beginning in June 1997 and continuing through July 1999, trained interviewers screened consecutive patients admitted to the ICUs to determine if they were eligible for participation. Eligible patients, or their proxies (usually next of kin) in the event that the patient could not respond, were asked to provide informed consent.

All patients receiving MV were screened, and patients receiving MV for at least 48 h were included. For patients with multiple admissions to the ICU, the first ICU admission with MV for > 48 h was used as the index admission. Exclusion criteria were patients hospitalized following solid-organ transplantation, subjects < 18 years of age, non-English-speaking persons, prisoners, and persons who were chronically ventilator-dependent at home or at a long-term ventilator facility. Patients who were transferred from other hospitals after intubation for > 24 h prior to admission to the UPMC were also excluded. In addition, recruitment was greater than expected several months into the study, so 10% of eligible participants were randomly excluded by not recruiting persons whose Social Security numbers had a terminal digit of “9.” There were no patients with burns in the group, as UPMC does not have a burn unit. The University of Pittsburgh Institutional Review Board approved the QOL-MV study protocol.

Data Collection

Trained research assistants screened the ICUs daily and identified patients receiving MV. Patient medical records were reviewed and eligibility status ascertained. Eligible subjects or family member proxies participated in a baseline interview in the hospital (during or after discharge from the ICU), and follow-up interviews occurred 2 months after the participants had been receiving MV for 48 h. Baseline data were collected by hospital chart review if neither the patient nor a proxy could be interviewed. Follow-up interviews were conducted in person whenever possible. Follow-up interviews were conducted by telephone if the patient lived outside a 50-mile radius or if the patient preferred a telephone interview. If neither the patient nor the proxy could participate in a full follow-up interview, a“ mini follow-up” interview was conducted in which only the patient’s vital status, current residence, and employment status were collected. Interviewers were trained to administer the forms by study investigators and were required to successfully administer practice interviews before starting independent data collection.

The following data were collected during baseline interviews:

1. Demographic variables, including age, sex, race, education, income, employment, and occupation.

2. Medical data, including admitting diagnoses to the ICU and comorbidities. The Charlson comorbidity score was used to assess comorbidity.13The Charlson comorbidity score is a weighted index that takes in to consideration specific diseases and their severity. A higher score is indicative of poor prognosis. Functional status was evaluated with Katz Activities of Daily Living (ADL), Instrumental Activities of Daily Living (IADL), and the physical functioning part of the Medical Outcomes Study Short-Form 36 (SF-36).1416 ADL assesses the need for help in performing basic activities of daily living, such as bathing, dressing, using the toilet, transfer in and out of bed or chair, continence, and feeding. IADL assesses a higher level function, such as ability to perform household chores and shopping. A higher score for ADL and IADL is indicative of more dependence. The physical function part of SF-36 evaluates ability to perform various physical activities, such as walking and climbing stairs. A higher score on SF-36 indicates better ability to perform these activities.

3. Length of stay in the ICU and hospital, duration of MV, need for tracheostomy, and resuscitation status. Length of stay of patients who needed MV within 24 h of extubation, or who were readmitted to an ICU within 24 h discharge from the ICU, was considered to be part of the same episode of MV and ICU admission. APACHE (acute physiology and chronic health evaluation) III score17 and mortality probability in the ICU and hospital were obtained from a hospital database.

4. Hospital charges obtained from the hospital financial database, and total hospital cost calculated by multiplying department charges by Medicare cost/charge ratio for individual departments. Cost per day was calculated by dividing the total hospital cost by hospital length of stay in days. Cost per hospital survivor and per 2-month survivor was calculated by dividing total hospital costs for the group by the number of patients discharged alive from the hospital, and the number of survivors at 2 months, respectively.10

The following data were collected during follow-up interviews: socioeconomic status (ie, income, education level attained), religiosity, social support, memory of the ICU experience, patient preference for undergoing MV, functional status data (ADL, IADL, and SF-36), and depressive symptoms using the Center for Epidemiologic Studies Depression Scale (CES-D).18CES-D evaluates depressive symptomatology, and a score > 16 is used as a screen for presence of depression. The individual domains for SF-36 were combined into physical (physical functioning, role- physical, bodily pain, general health) and mental (vitality, social functioning, role-emotional, mental health) components. Caregivers, who were identified by the patients or the person identifying himself or herself as the caregiver, provided data on caregiver burden19 and involvement with patient care. The interviews with caregivers were done at the same time as the interviews with patients, and were performed in person or by phone based on subject preference or distance from UPMC. Vital status information was collected by prospectively following up the subjects. Dates of death were collected through interviews with family members of deceased subjects. Date of death information was verified, and cause of death was ascertained by reviewing death certificates. For any patient who was unavailable for follow-up, we performed a minimum of two searches of death certificates in his or her state of residence. The final death certificate search occurred at least 1 year following the last date of contact. If no death certificate was found after these searches, the patient was assumed to have survived. The analysis to identify factors related to 2-month mortality was performed both including and excluding those subjects who were unavailable for follow-up, in order to test the effect, if any, of classifying the unavailable subjects into the survival group.

Statistical Analysis

The QOL-MV sample is described using measures of central tendency and variability for continuous variables and frequency distributions for categorical variables. Statistics are presented overall and for the 2-month nonsurvivor group.

The relations between patient factors and 2-month mortality were examined through logistic regression analyses. The outcome (dependent) variable in the logistic regression analysis was whether or not the patient survived for 2 months following MV. A multiple logistic regression model was fit to identify independent predictors of 2-month mortality. The following covariates were included in the model: age (measured on a continuous scale with the unit of measurement equal to 1 decade), sex (female subjects–reference group), race (white subjects–reference group, nonwhites), Acute Physiology Score (APS) on the enrollment day (continuous, unit of measurement was a 10-point score change), Charlson comorbidity scores (continuous), type of patient (medical–reference group, surgical, or trauma patient), and IADL scores (score = 0–reference group, score ≥ 1, missing score). A missing score on the IADL scale typically indicated that the patient was unable to participate in the baseline interview due to severe physical or mental impairment, or that the patient died prior to contact with the family. In addition to the main effects model, all pairwise interactions were examined in a stepwise model-building fashion. Entry and exit criterion for the pairwise interactions was statistical significance of 0.05. The Hosmer-Lemeshow goodness-of-fit statistic was calculated.20Also, the area under the receiver operating characteristic (ROC) curve was computed to determine the predictive properties of the final model.21

Two-month functional status measures are reported for those that survived to the 2-month follow-up interview. Data are presented separately by respondent (patient or proxy) because many of the 2-month follow-up interviews were conducted with a family-member surrogate due to the patient’s inability to complete the interview.

Study Sample and Baseline Characteristics

During the recruitment period, 2,237 patients received 2,557 episodes of MV in the study ICUs (Fig 1 ). Of the 2,237 patients, 1,123 patients (50.2%) received MV for ≥ 48 h. Two hundred twenty-nine patients were excluded because they had pretransplantation or posttransplantation status, were chronically ventilator dependent, were intubated for 24 h prior to ICU admission, or did not speak English. Of the remaining 894 patients, 77 patients or their families refused to participate in the study. Thus, the final sample is based on 817 subjects. The 2-month mortality rate was 43%. Among the 466 survivors, we were unable to interview 98 survivors (21%) at 2 months. Eighty-three of these 98 subjects were known to be alive at 2 months since we were able to interview them at later follow-up interviews (ie, 6 months or 12 months), while 15 patients were unavailable for follow-up at 2 months. As we were unable to find any evidence of these 15 subjects being deceased through two death certificate searches, we assumed that they were alive at 2 months. There were no statistically significant differences between these 15 subjects and the 802 patients we were able to successfully follow up in terms of age, sex, race, education, APACHE III scores on the first ICU day, IADL, and SF-36 physical functioning. Patient recruitment is described in Figure 1.

Sociodemographic and clinical characteristics of the QOL-MV cohort are presented in Tables 1 , 2. The median and mean ages of the cohort were 65 years and 60.4 years, respectively, with a SD of 18.8 years and a range of 18 to 96 years. There were slightly fewer women (46%) than men. The QOL-MV cohort was 84.5% white and 15.2% African American. Eighty-one percent of the patients completed high school, and 33% had some post-high school education. The median income was approximately $20,000/yr. Fifty-one percent of the cohort were married.

Eighty-nine percent of the patients were at home prior to the hospitalization resulting in their enrollment in the study, with nearly all of the other participants being admitted from nursing homes (8.5%) or rehabilitation facilities (1.6%). Sixty-six percent of the patients were retired or not employed due to some other reason including disability. Twenty-three percent were employed and another 10% were homemakers. Sixty-eight percent of the cohort was admitted for a medical condition (mostly respiratory). Twenty-eight percent of the patients had a history of chronic pulmonary disease. The median and mean ± SD APACHE III scores on the day of admission to the ICU were 64 and 68.0 ± 28.3, respectively. The mean APACHE III scores on the QOL-MV enrollment day (typically the third ICU day) was 63.0 ± 27.6, with a median of 59. The corresponding median and mean ± SD expected probabilities of death in the hospital predicted by the APACHE system were 0. 31 and 0.37 ± 0.28 for the day of ICU admission, and 0.42 and 0.46 ± 0.31 for enrollment day, respectively. Median length of stay in the ICU and in the hospital were 11 days and 20 days, respectively. The median time receiving MV was 8.6 days. Thirty-four percent of the patients underwent tracheostomy.

Median total hospital cost and daily cost were $56,100 (25th to 75th percentile, $31,900 to $83,400), and $2,700 (25th to 75th percentile, $2,200 to $3,400, respectively). Costs per hospital survivor and 2-month survivor were $96,223 and $116,872, respectively. One hundred ten hospital survivors (21%) were discharged to home, 190 survivors (36.1%) were discharged to rehabilitation facilities, 144 survivors (27.4%) went to nursing homes, 48 survivors (5.9%) went to chronic ventilator facilities, and the remaining 34 survivors (6.5%) were discharged to other hospitals.

2-Month Mortality: Univariate Analysis

The 2-month mortality rate was 43.0% (351 of 817 patients) when counting the 98 patients unavailable for follow-up at 2 months (for whom two searches yielded no death certificates) as survivors. Nonsurvivors tended to be older than survivors (odds of dying within 2 months were 31% higher for every decade increase in age). There was not a significant difference in 2-month survival rates by sex. Though 2-month mortality was slightly higher among nonwhites compared to whites, the difference was not statistically significant (p = 0.10). Mortality at 2 months did not differ significantly by education, income, or marital status. Two-month mortality was higher in patients admitted from a nursing home, rehabilitation, or other facility compared to patients admitted from home (55.7% vs 41.4%, p = 0.01). Medical patients had significantly higher mortality than those admitted for surgery or trauma. All measures of functional status and patient acuity we examined were related to 2-month mortality. For every additional comorbidity, as assessed by the Charlson comorbidity score, the odds of dying by 2 months increased by 24% (95% confidence interval [CI] for odds ratio [OR], 1.17 to 1.32). The ORs for mortality for each additional ADL and IADL requiring assistance were 1.24 (95% CI, 1.15 to 1.34) and 1.23 (95% CI, 1.16 to 1.31), respectively. Higher SF-36 physical functioning scores, implying better physical well-being, were related to a decreased odds of 2-month mortality (OR, 0.85; 95% CI, 0.81 to 0.89). Measures from the APACHE III system were significantly associated with 2-month mortality. For example, the odds of dying by 2 months increased by 34% for each 10-U increase in the ICU day 1 APACHE score. Those with longer lengths of stay in the hospital and ICU and longer time receiving MV had significantly reduced odds of 2-month mortality (Table 3 ) because those who died within 2 months were likely to die soon after they received MV for 48 h. Increased total hospital costs were inversely associated with odds of 2-month mortality, as 2-month mortality decreased by 9% for every $10,000 increase in total hospital costs, likely due to the strong relationship between hospital length of stay and total costs. However, average daily hospital costs were positively associated with odds of 2-month mortality (OR for each $1,000 increase, 1.52; 95% CI, 1.32 to 1.76), implying greater intensity of resource utilization for those who ultimately died.

Multivariable Analysis

Because we found comparable results between the analysis that included the 15 subjects who were unavailable for follow-up (and assumed alive at 2 months) and the analysis that excluded these subjects, we report only the results on the former. Multiple logistic regression analysis revealed that after adjusting for other demographic and clinical measures, age was significantly related to mortality 2 months after MV. For each decade increase in age, the odds of dying within 2 months increased by 34% (95% CI, 17% to 54%; p = 0.0001). APS (enrollment day) was also independently related to dying by 2 months (OR, 1.48 for a 10-point increase in APS; 95% CI, 1.36 to 1.62; p = 0.0001). Those who had more chronic disease at baseline (as assessed by the Charlson comorbidity scores) were at higher risk of dying by 2 months (OR, 1.88; 95% CI, 1.35 to 2.62; p = 0.0002). The odds of dying by 2 months for subjects with IADL scores ≥ 1 were 63% higher than that for subjects with IADL scores of zero. Patients with missing IADL scores (primarily patients with physical or mental conditions that prevented their participation in the baseline interview or who had died prior to contact with the family) were at nearly five times the risk of dying by the 2-month follow-up than those with an IADL score of zero. Age and Charlson comorbidity score was the only interaction that entered (and remained) in the regression model. The interaction was such that mortality among older subjects with greater comorbidity was less than multiplicative. Sex, race, and diagnostic group (medical, surgical, or trauma) were not significantly related to dying by 2 months (p > 0.05). The Hosmer-Lemeshow statistic had a p value of 0.10, implying that there was no evidence of lack of model fit. The model had good predictive properties, as the area under the ROC curve was 0.82. Detailed results of the regression analysis are presented in Table 4 .

Functional Status at 2-Month Follow-up

Of the 466 patients alive at 2 months, full follow-up interviews were conducted on 232 patients, including 134 patient interviews and 98 proxy interviews. Only vital status, employment status, and residence data (mini follow-up) were collected on 136 patients, while only vital status was known for 83 subjects with missed interviews at 2 months. The other 15 patients were those who were unavailable for follow-up, and presumed to be alive. Sixty-three percent of the patients (217 of 344 patients) for whom data on residence at 2 months were available were residing at home at the 2-month follow-up interview (Table 5 ). Most of the 134 patients (87%) who were able to do the full follow-up interview were at home. Fifty-three percent of 98 patients for whom proxy interviews were conducted resided at home at the time of the 2-month follow-up interview, 26% were in a nursing home, and 15% were at a rehabilitation facility. Of the 136 subjects with mini follow-up, 45% were residing at home at 2 months, 17% were at a rehabilitation facility, and 26% were at a nursing home.

Employment data prior to hospitalization were available for 772 patients (94%), 176 of whom (22.8%) were employed. Forty-six of these 176 patients (26.1%) had died by 2 months after study entry, and 32 of the 2-month survivors (9.4%) returned to work. Many of the 2-month survivors were dependent on others to perform basic tasks. Median ADL and IADL scores were 1 and 4, respectively, for patients who participated in the follow-up interview, and median ADL and IADL scores were 5 and 7, respectively, for patients who had a proxy participate in the interview. Median scores for the SF 36 physical component and mental components were 30 and 53, respectively, in patients who were able to participate in the full interview. Median and mean CES-D scores were 13 (25th to 75th percentile, 6 to 20) and 14.1 ± 9.6, respectively. Thirty-five percent of the patients had a CES-D score≥ 16, indicating that many patients are at risk for clinical depression. Fifty percent of the patients described their general health at the 2-month interview as good, very good, or excellent, and 16% stated that their current health status was better compared to a year before. Seventy-eight percent of the patients for whom caregiver data were available had a caregiver, and 17% were living alone without a caregiver.

Mortality

Patients receiving MV for any duration in an ICU have higher in-hospital mortality than patients who do not need MV support. The reported duration of MV prior to inclusion in different studies has varied from a few hours to 2 to 3 days. Hospital mortality was reported to vary from 50 to 67% in a review by Stauffer et al6 in 1993. Other studies5,1011,2223 report in-hospital mortality ranging from 23 to 43%. The differences in these studies may be due to different inclusion criteria resulting in different patient populations and different criteria for hospital discharge. Few of these studies report long-term mortality rates after discharge from the hospital.

Hospital outcomes are one measure of the effectiveness of therapeutic interventions. However, hospital outcomes are inadequate as the sole measures of therapeutic effectiveness. Evaluation of patient outcomes after hospital discharge, including functional status, morbidity, quality of life, return to work and usual activities, and mortality, are important as well. A vital issue for evaluating hospital services is how long after discharge patient outcomes should be assessed. Furthermore, hospital outcomes are influenced by discharge strategies. For example, hospitals that are able to discharge patients to other facilities (eg, rehabilitation, nursing homes) sooner than other hospitals will have shorter length of stay and lower in-hospital mortality. Hence, hospital mortality rates are not comparable across studies.

We report a 2-month mortality of 43%. We found that patients with medical diagnoses had a higher mortality rate than those with trauma or a surgical diagnosis. Kollef et al11 reported similar findings for hospital mortality. Our data show that those who did not survive at 2 months were on average more severely ill on entry into the ICU and into the QOL-MV study than those who did survive, as indicated by severity-of-illness and functional-status measures, and were on average older and more likely to have been admitted from a nursing home or rehabilitation facility than survivors. Nonsurvivors had lower average total hospital cost and shorter length of stay, indicating that nonsurvivors died earlier during hospitalization. Also, in this sample, every decade of increase in age increased the risk of death at 2 months by 34%. This association between age and outcome following prolonged MV was constant across the entire age range. This result is different from that reported by Ely et al,5 who found no difference in hospital mortality in older and younger patients, but they did not report post-hospital discharge mortality. Similar to other studies, we report that comorbidity and functional status prior to hospitalization were also risk factors for increased 2-month mortality.2425

2-Month Functional Status

These data show that functional status, as measured by ADL, IADL, or SF-36, was worse at 2 months compared to patients’ prehospital levels. Additionally, SF-36 scores were lower than those reported for the general population.16 Similar decline in function was reported after critical illness in elderly patients, patients requiring intensive care for a prolonged period, and patients with ARDS and lung injury.2630 Engoren et al31 studied survival and functional status of patients receiving MV support ≥ 7 days after cardiac surgery. The follow-up interval was variable. They reported 59% survival at 5 years, and 59% of the survivors described their health status as good. We reported more severe limitations in physical function compared to mental function in our cohort. This finding is similar to those reported by others2729 after acute lung injury or critical illness requiring intensive care. As functional status was worse at 2 months, longer-term follow-up is needed because recovery of functional abilities may take > 2 months after a critical illness. The functional status of patients where a proxy was the respondent was worse compared to the functional status of those patients who participated in the interview. As there is a significant difference in functional status based on whether the respondent was patient or proxy, one should be cautious in combing functional status data obtained from patient and proxy interviews. The mean CES-D score for those who were able to complete the 2-month interview was higher than that reported for general population and the CES-D score at 1 month after discharge from the hospital found in a group of elderly survivors of critical illness.12,18 The prevalence of significant depressive symptomatology (≥ 16 on CES-D) was also higher than the 15 to 16% reported for older adults in the general population, and indicates the need for assessment for depression and treatment.3233 Although the objective measures of functional status were worse, 50% of the patients in this study described their quality of life as good or better, and similar results were reported by others.3435 Patient’s scores on the quality-of-life assessments differ from their scores on objective scales because quality of life is not “a description of patient’s health status” but rather “the way that patients perceive and react to their health status.”35

Not surprisingly, a large percentage (78%) of these patients had an informal caregiver providing support 2 months after entry into the study. The decline in functional status, increased prevalence of depressive symptoms, and involvement of a caregiver indicate that patient as well as caregiver evaluations are warranted after hospital discharge. These evaluations should be made to determine the health needs of the survivors of prolonged MV, and the impact on their caregivers, so that appropriate health and supportive services can be provided to both. It is very likely that the patients receive at least some follow-up services, but studies have shown that it is much less likely that their caregivers receive adequate assessments or services.36

Hospital Costs

Hospital costs for the patients were relatively high and costs for nonsurvivors were slightly lower compared to survivors (median cost, $47,100 vs $60,800 in nonsurvivors and survivors, respectively.) Median cost per day was slightly higher for nonsurvivors compared to survivors ($2,900/d and $2,600/d, respectively), but the differences in total hospital cost between nonsurvivors and survivors were mostly related to the increased hospital length of stay for survivors (16 days for nonsurvivors vs 23 days for survivors). Gracey et al37reported hospital costs for elderly patients aged ≥ 65 years and receiving MV for ≥ 48 h as $31,896 in 1985, which is comparable to that reported by us after adjusting for health-care inflation between 1985 and 1997.38 Ely and coworkers5 reported hospital costs of patients receiving MV in different age groups (< 75 years of age and ≥ 75 years of age) as $29,049 and $21,292 in the younger group and older group, respectively. The lower costs in the study by Ely and colleagues5 are probably related to the patient population, which included only medical and coronary ICUs patients.

Costs per hospital and 2-month survivor in the current study were $96,223 and $116,872, respectively. The cost per survivor that we report is higher than that reported by Kurek et al10 for patients receiving MV under diagnosis-related group 475 in New York State in 1993. The difference is probably related to the entry criterion, which was all patients receiving MV under diagnosis-related group 475 in the study by Kurek et al,10 whereas the current study included only patients receiving MV for ≥48 h. In addition, we calculated costs from the hospital charges, whereas Kurek et al10 used the hospital reimbursement, which may be lower than the actual costs.

Limitations of this study include missing data on functional status, lack of physiologic data specific to the acute illness, and incomplete cost data. We obtained information on functional status at base line for 75 to 85% of the patients. It was difficult to obtain IADL and SF-36 data by chart review, and the missing data were mostly from patients who died before we could interview the patient/family. As our primary focus was long-term mortality and functional status and we had limited resources, we focused our data collection toward long-term follow-up and did not collect detailed physiologic information, except for APACHE III score, while the patients were in the ICU. We obtained follow-up functional status information on 79% of the survivors, and similar follow-up rates were reported by others.26,29 The functional status was based on data obtained by proxy interview, as many of the patients were not able to participate in the interview at 2 months. We did not obtain physician costs and costs after discharge from the hospital. So, our cost data underestimate total costs and cost per 2-month survivor.

This prospective study of a large cohort of patients receiving MV demonstrates that age and functional status prior to hospitalization are factors predictive of mortality at 2 months. Such information may be useful in analyzing future trials of treatment effectiveness and in discussing goals of therapy for patients receiving prolonged MV. In addition, since family and caregiver support, as well as the patient functional and psychological status, play a major role in survivor quality-of-life issues, the caregiver support structure and its impact on long-term outcome needs to be evaluated further. Improved caregiver support may prove to be a highly effective and efficient mechanism for improving long-term psychological and functional outcomes in patients receiving prolonged MV.

The QOL-MV Study Investigators are Lakshmipathi Chelluri, MD, MPH, FCCP, Armando J. Rotondi, PhD, Carl A. Sirio, MD, Michael P. Donahoe, MD, and Micheal R. Pinsky, MD, University of Pittsburgh School of Medicine; Aaron B. Mendelsohn, PhD, and Steven H. Belle, PhD, Graduate School of Public Health; and Richard Schulz, PhD, University Center for Social and Urban Research, University of Pittsburgh, Pittsburgh, PA.

Abbreviations: ADL = Activities of Daily Living; APACHE = acute physiology and chronic health evaluation; APS = acute physiology score; CES-D = Center for Epidemiologic Studies Depression Scale; CI = confidence interval; IADL = Instrumental Activities of Daily Living Deficits; MV = mechanical ventilation; OR = odds ratio; QOL-MV = Quality of Life after Mechanical Ventilation in the Aged; ROC = receiver operating characteristic; SF-36 = Medical Outcomes Study Short-Form 36; UPMC = University of Pittsburgh Medical Center

Supported by National Institute On Aging grant No. RO1AG11979.

Figure Jump LinkFigure 1. Flow diagram of patient enrollment and follow-up.Grahic Jump Location
Table Graphic Jump Location
Table 1. ORs of 2-Month Mortality by Sociodemographic Characteristics Among Patients in the QOL-MV Study*
* 

Data are presented as No. (%) unless otherwise indicated. The first group listed in column 1 (female, white, less than high school, married, < $20,000, home, employed) is the reference group.

 

Increase in odds of 2-month mortality for each decade increase in age.

 

Comparisons are home vs not home, and employed vs not employed, respectively.

Table Graphic Jump Location
Table 2. ORs of 2-Month Mortality By Clinical Characteristics Among Patients in the QOL-MV Study*
* 

Data are presented as median (25th, 75th percentile) unless otherwise indicated.

 

Reference group is medical.

 

Increase in odds of 2-month mortality for each 1-U increase.

§ 

Increase in odds of 2-month mortality for each 10-U increase.

 

Increase in odds of 2-month mortality for each 0.10-U increase.

Table Graphic Jump Location
Table 3. ORs of 2-Month Mortality by Resource Utilization Among Patients Receiving Prolonged MV in the QOL-MV Study*
* 

Data are presented as median (25th, 75th percentile) unless otherwise indicated.

 

Increase in odds of 2-month mortality for each $10,000 increase.

 

Increase in odds of 2-month mortality for each $1,000 increase.

Table Graphic Jump Location
Table 4. Results of Multiple Logistic Regression Analysis of 2-Month Mortality Among Patients in the QOL-MV Study (n = 759)*
* 

Model predicting the probability that patient died by 2-months follow-up; area under the ROC curve = 0.82. Hosmer-Lemeshow goodness-of-fit statistic = 13.46 with eight degrees of freedom, p = 0.10.

 

Based on a 10-U increase (eg, decades).

 

Medical is the reference group.

Table Graphic Jump Location
Table 5. Functional Status at 2 Months Among Survivors Enrolled in the QOL-MV Study*
* 

Data are presented as median (25th, 75th percentile) unless otherwise indicated.

 

Only vital status, residence, and current employment status data collected.

Older American 2000: key indicators of well-being. Federal Interagency Forum on Aging-Related Statistics. Available at: www.agingstats.go. Accessed January 14, 2002.
 
Esteban, A, Anzueto, A, Alia, I, et al How is mechanical ventilation employed in the intensive care unit? An international utilization review.Am J Respir Crit Care Med2000;161,1450-1458. [PubMed]
 
Suresh, R, Kupfer, YY, Tessler, S The graying of the intensive care unit: demographic changes 1988–98 [abstract]. Crit Care Med. 1999;;27 ,.:A27. [PubMed]
 
Wagner, DP Economics of prolonged mechanical ventilation.Am Rev Respir Dis1989;140,S14-S18. [PubMed]
 
Ely, WE, Evans, GW, Haponik, EF Mechanical ventilation in a cohort of elderly patients admitted to an intensive care unit.Ann Intern Med1999;131,96-104. [PubMed]
 
Stauffer, JL, Fayter, NA, Graves, B Survival following mechanical ventilation for acute respiratory failure in adult men.Chest1993;104,1222-1229. [PubMed] [CrossRef]
 
Cohen, IL, Lambrinos, J, Fein, IA Mechanical ventilation for the elderly patient in Intensive care: incremental charges and benefits.JAMA1993;269,1025-1029. [PubMed]
 
Swinburne, AJ, Fedullo, AJ, Bixby, K, et al Respiratory failure in the elderly: analysis of outcome after treatment with mechanical ventilation.Arch Intern Med1993;153,1657-1662. [PubMed]
 
Chelluri, L, Grenvik, A, Silverman, M Intensive care for critically ill elderly: mortality, costs and quality of life.Arch Intern Med1995;155,1013-1022. [PubMed]
 
Kurek, CJ, Dewar, D, Lambrinos, J, et al Clinical and economic outcome of mechanically ventilated patients in New York State during 1993: analysis of 10,473 cases under DRG 475.Chest1998;114,214-222. [PubMed]
 
Kollef, MH, O’Brien, JD, Silver, P The impact of gender on outcome from mechanical ventilation.Chest1997;111,434-441. [PubMed]
 
Chelluri, L, Pinsky, MR, Donahoe, MP, et al Long-term outcome of critically ill elderly patients requiring intensive care.JAMA1993;269,3119-3123. [PubMed]
 
Charlson, ME, Pompei, P, Ales, KL A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.J Chronic Dis1987;40,373-383. [PubMed]
 
Katz, S, Ford, AB, Moskowitz, RW Studies of illness in the aged: the index of ADL; a standardized measure of biological and psychosocial function.JAMA1963;185,914-919. [PubMed]
 
Lawton, MP, Brody, EM Assessment of older people: self-maintaining and instrumental activities of daily living.Gerontologist1969;9,179-186. [PubMed]
 
Ware J. SF 36 health survey. Boston, MA: Health Institute, New England Medical Center, 1993.
 
Knaus, WA, Wagner, DP, Draper, E, et al The APACHE III prognostic system: risk prediction of hospital mortality for critically ill hospitalized adults.Chest1991;100,1619-1636. [PubMed]
 
Radloff, L The Center for Epidemiological Studies depression scale: a self-report depression scale for research in the general population.Appl Psychol Meas1977;1,385-401
 
Teri, L, Traux, P, Logsdon, R, et al Assessment of behavioral problems in dementia: the revised memory and behavior problems checklist.Psychol Aging1992;7,622-631. [PubMed]
 
Hosmer, DW, Lemeshow, S. Applied logistic regression. 1989; John Wiley and Sons. New York, NY:.
 
Hanley, JA, McNeil, BJ The meaning and use of the area under a receiver operating characteristic (ROC) curve.Radiology1982;143,29-36. [PubMed]
 
Kollef, MH, Ahrens, TS, Shannon, W Clinical predictors and outcome for patients requiring tracheotomy in the intensive care unit.Crit Care Med1999;27,1714-1720. [PubMed]
 
Beherendt, CE Acute respiratory failure in the United States: incidence and 31-day survival.Chest2000;118,1100-1105. [PubMed]
 
Narain, P, Reubenstein, LZ, Wieland, DG, et al Predictors of immediate and 6-month outcomes in hospitalized elderly patients: the importance of functional status.J Am Geriatr Soc1988;36,775-783. [PubMed]
 
Inouye, SK, Peduzzi, PN, Robison, JT, et al Importance of functional measures in predicting mortality among older hospitalized patients.JAMA1998;279,1187-1193. [PubMed]
 
Ridley, SA, Chrispin, PS, Scotton, H, et al Changes in quality of life after intensive care: comparison with normal data.Anaesthesia1997;52,195-202. [PubMed]
 
Weinert, CR, Gross, CR, Kamgas, JR, et al Health-related quality of life after acute lung injury.Am J Respir Crit Care Med1997;156,1120-1128. [PubMed]
 
Schelling, G, Stoll, C, Haller, M, et al Health-related quality of life and posttraumatic stress disorder in survivors of the acute respiratory distress syndrome.Crit Care Med1998;26,651-659. [PubMed]
 
Niskanen M, Ruokonen, Takala J, et al. Quality of life after prolonged intensive care. Crit Care Med 1999; 27:1132–1139.
 
Davidson, TA, Caldwell, ES, Curtis, JR, et al Reduced quality of life in survivors of acute respiratory distress syndrome compared with critically ill control patients.JAMA1999;281,354-360. [PubMed]
 
Engoren, M, Buderer, NF, Zacharias, A Long-term survival and health status after prolonged mechanical ventilation after cardiac surgery.Crit Care Med2000;28,2742-2749. [PubMed]
 
Huntley J, Brock DB, Ostfeld AM, et al. Established population for epidemiologic studies of the elderly (resource data book). Washington, DC: National Institute on Aging, 1986; National Institutes of Health publication No. 86–2443.
 
Callahan, MC, Frederic, DW The effect of gender and race on the measurement properties of the CES-D in older adults.Med Care1994;32,341-356. [PubMed]
 
Tsevat, J, Dawson, NV, Wu, AW, et al Health values of hospitalized patients 80 years or older.JAMA1998;279,371-375. [PubMed]
 
Gill, TM, Feinstein, AR A critical appraisal of the quality of quality of life measurements.JAMA1994;272,619-626. [PubMed]
 
Given, CW, Given, B, Stommel, M, et al The caregiver reaction assessment (CRA) for caregivers to persons with chronic physical and mental impairments.Res Nurs Health1992;15,271-283. [PubMed]
 
Gracey, DR, Gillespie, D, Nobrega, F, et al Financial implications of prolonged ventilator care of Medicare patients under the prospective payment system: a multicenter study.Chest1987;91,424-427. [PubMed]
 
Braden, BR, Cowan, CA, Lazenby, HC, et al National health expenditures, 1997.Health Care Financ Rev1998;20,83-121. [PubMed]
 

Figures

Figure Jump LinkFigure 1. Flow diagram of patient enrollment and follow-up.Grahic Jump Location

Tables

Table Graphic Jump Location
Table 1. ORs of 2-Month Mortality by Sociodemographic Characteristics Among Patients in the QOL-MV Study*
* 

Data are presented as No. (%) unless otherwise indicated. The first group listed in column 1 (female, white, less than high school, married, < $20,000, home, employed) is the reference group.

 

Increase in odds of 2-month mortality for each decade increase in age.

 

Comparisons are home vs not home, and employed vs not employed, respectively.

Table Graphic Jump Location
Table 2. ORs of 2-Month Mortality By Clinical Characteristics Among Patients in the QOL-MV Study*
* 

Data are presented as median (25th, 75th percentile) unless otherwise indicated.

 

Reference group is medical.

 

Increase in odds of 2-month mortality for each 1-U increase.

§ 

Increase in odds of 2-month mortality for each 10-U increase.

 

Increase in odds of 2-month mortality for each 0.10-U increase.

Table Graphic Jump Location
Table 3. ORs of 2-Month Mortality by Resource Utilization Among Patients Receiving Prolonged MV in the QOL-MV Study*
* 

Data are presented as median (25th, 75th percentile) unless otherwise indicated.

 

Increase in odds of 2-month mortality for each $10,000 increase.

 

Increase in odds of 2-month mortality for each $1,000 increase.

Table Graphic Jump Location
Table 4. Results of Multiple Logistic Regression Analysis of 2-Month Mortality Among Patients in the QOL-MV Study (n = 759)*
* 

Model predicting the probability that patient died by 2-months follow-up; area under the ROC curve = 0.82. Hosmer-Lemeshow goodness-of-fit statistic = 13.46 with eight degrees of freedom, p = 0.10.

 

Based on a 10-U increase (eg, decades).

 

Medical is the reference group.

Table Graphic Jump Location
Table 5. Functional Status at 2 Months Among Survivors Enrolled in the QOL-MV Study*
* 

Data are presented as median (25th, 75th percentile) unless otherwise indicated.

 

Only vital status, residence, and current employment status data collected.

References

Older American 2000: key indicators of well-being. Federal Interagency Forum on Aging-Related Statistics. Available at: www.agingstats.go. Accessed January 14, 2002.
 
Esteban, A, Anzueto, A, Alia, I, et al How is mechanical ventilation employed in the intensive care unit? An international utilization review.Am J Respir Crit Care Med2000;161,1450-1458. [PubMed]
 
Suresh, R, Kupfer, YY, Tessler, S The graying of the intensive care unit: demographic changes 1988–98 [abstract]. Crit Care Med. 1999;;27 ,.:A27. [PubMed]
 
Wagner, DP Economics of prolonged mechanical ventilation.Am Rev Respir Dis1989;140,S14-S18. [PubMed]
 
Ely, WE, Evans, GW, Haponik, EF Mechanical ventilation in a cohort of elderly patients admitted to an intensive care unit.Ann Intern Med1999;131,96-104. [PubMed]
 
Stauffer, JL, Fayter, NA, Graves, B Survival following mechanical ventilation for acute respiratory failure in adult men.Chest1993;104,1222-1229. [PubMed] [CrossRef]
 
Cohen, IL, Lambrinos, J, Fein, IA Mechanical ventilation for the elderly patient in Intensive care: incremental charges and benefits.JAMA1993;269,1025-1029. [PubMed]
 
Swinburne, AJ, Fedullo, AJ, Bixby, K, et al Respiratory failure in the elderly: analysis of outcome after treatment with mechanical ventilation.Arch Intern Med1993;153,1657-1662. [PubMed]
 
Chelluri, L, Grenvik, A, Silverman, M Intensive care for critically ill elderly: mortality, costs and quality of life.Arch Intern Med1995;155,1013-1022. [PubMed]
 
Kurek, CJ, Dewar, D, Lambrinos, J, et al Clinical and economic outcome of mechanically ventilated patients in New York State during 1993: analysis of 10,473 cases under DRG 475.Chest1998;114,214-222. [PubMed]
 
Kollef, MH, O’Brien, JD, Silver, P The impact of gender on outcome from mechanical ventilation.Chest1997;111,434-441. [PubMed]
 
Chelluri, L, Pinsky, MR, Donahoe, MP, et al Long-term outcome of critically ill elderly patients requiring intensive care.JAMA1993;269,3119-3123. [PubMed]
 
Charlson, ME, Pompei, P, Ales, KL A new method of classifying prognostic comorbidity in longitudinal studies: development and validation.J Chronic Dis1987;40,373-383. [PubMed]
 
Katz, S, Ford, AB, Moskowitz, RW Studies of illness in the aged: the index of ADL; a standardized measure of biological and psychosocial function.JAMA1963;185,914-919. [PubMed]
 
Lawton, MP, Brody, EM Assessment of older people: self-maintaining and instrumental activities of daily living.Gerontologist1969;9,179-186. [PubMed]
 
Ware J. SF 36 health survey. Boston, MA: Health Institute, New England Medical Center, 1993.
 
Knaus, WA, Wagner, DP, Draper, E, et al The APACHE III prognostic system: risk prediction of hospital mortality for critically ill hospitalized adults.Chest1991;100,1619-1636. [PubMed]
 
Radloff, L The Center for Epidemiological Studies depression scale: a self-report depression scale for research in the general population.Appl Psychol Meas1977;1,385-401
 
Teri, L, Traux, P, Logsdon, R, et al Assessment of behavioral problems in dementia: the revised memory and behavior problems checklist.Psychol Aging1992;7,622-631. [PubMed]
 
Hosmer, DW, Lemeshow, S. Applied logistic regression. 1989; John Wiley and Sons. New York, NY:.
 
Hanley, JA, McNeil, BJ The meaning and use of the area under a receiver operating characteristic (ROC) curve.Radiology1982;143,29-36. [PubMed]
 
Kollef, MH, Ahrens, TS, Shannon, W Clinical predictors and outcome for patients requiring tracheotomy in the intensive care unit.Crit Care Med1999;27,1714-1720. [PubMed]
 
Beherendt, CE Acute respiratory failure in the United States: incidence and 31-day survival.Chest2000;118,1100-1105. [PubMed]
 
Narain, P, Reubenstein, LZ, Wieland, DG, et al Predictors of immediate and 6-month outcomes in hospitalized elderly patients: the importance of functional status.J Am Geriatr Soc1988;36,775-783. [PubMed]
 
Inouye, SK, Peduzzi, PN, Robison, JT, et al Importance of functional measures in predicting mortality among older hospitalized patients.JAMA1998;279,1187-1193. [PubMed]
 
Ridley, SA, Chrispin, PS, Scotton, H, et al Changes in quality of life after intensive care: comparison with normal data.Anaesthesia1997;52,195-202. [PubMed]
 
Weinert, CR, Gross, CR, Kamgas, JR, et al Health-related quality of life after acute lung injury.Am J Respir Crit Care Med1997;156,1120-1128. [PubMed]
 
Schelling, G, Stoll, C, Haller, M, et al Health-related quality of life and posttraumatic stress disorder in survivors of the acute respiratory distress syndrome.Crit Care Med1998;26,651-659. [PubMed]
 
Niskanen M, Ruokonen, Takala J, et al. Quality of life after prolonged intensive care. Crit Care Med 1999; 27:1132–1139.
 
Davidson, TA, Caldwell, ES, Curtis, JR, et al Reduced quality of life in survivors of acute respiratory distress syndrome compared with critically ill control patients.JAMA1999;281,354-360. [PubMed]
 
Engoren, M, Buderer, NF, Zacharias, A Long-term survival and health status after prolonged mechanical ventilation after cardiac surgery.Crit Care Med2000;28,2742-2749. [PubMed]
 
Huntley J, Brock DB, Ostfeld AM, et al. Established population for epidemiologic studies of the elderly (resource data book). Washington, DC: National Institute on Aging, 1986; National Institutes of Health publication No. 86–2443.
 
Callahan, MC, Frederic, DW The effect of gender and race on the measurement properties of the CES-D in older adults.Med Care1994;32,341-356. [PubMed]
 
Tsevat, J, Dawson, NV, Wu, AW, et al Health values of hospitalized patients 80 years or older.JAMA1998;279,371-375. [PubMed]
 
Gill, TM, Feinstein, AR A critical appraisal of the quality of quality of life measurements.JAMA1994;272,619-626. [PubMed]
 
Given, CW, Given, B, Stommel, M, et al The caregiver reaction assessment (CRA) for caregivers to persons with chronic physical and mental impairments.Res Nurs Health1992;15,271-283. [PubMed]
 
Gracey, DR, Gillespie, D, Nobrega, F, et al Financial implications of prolonged ventilator care of Medicare patients under the prospective payment system: a multicenter study.Chest1987;91,424-427. [PubMed]
 
Braden, BR, Cowan, CA, Lazenby, HC, et al National health expenditures, 1997.Health Care Financ Rev1998;20,83-121. [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543