The Minimum Data Set for Nursing Home Activities of Daily Living (MDS-ADL) includes information about mobility in 7 activities: mobility while in bed, transferring, ambulation, dressing, eating, toileting, and personal hygiene. Performance of each of these ADLs is rated using a scoring system of 0 to 4 points, ranging from 0, indicating independence, to 4, indicating total dependence. Possible total scores range from 0, for independence in all ADLs, to 28, indicating total dependence in all ADLs.
eAppendix. List of ICD-9 Diagnosis and Procedure Codes for Breast Cancer and Breast Cancer Surgery
eTable 1. 30-day Mortality Rates According to Procedure Type
eTable 2. 1-Year Mortality Rates According to Procedure Type
eTable 3. 1-Year Functional Decline Rates According to Procedure Type
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Tang V, Zhao S, Boscardin J, et al. Functional Status and Survival After Breast Cancer Surgery in Nursing Home Residents. JAMA Surg. 2018;153(12):1090–1096. doi:10.1001/jamasurg.2018.2736
What are long-term functional and mortality outcomes among older female residents of nursing homes after breast cancer surgery?
In this study of 5969 female nursing home residents, the highest 30-day and 1-year mortality occurred after lumpectomy (least invasive procedure) and the greatest 1-year functional decline occurred after lumpectomy or mastectomy with axillary lymph node dissection (most invasive procedure). Among all patients who underwent breast cancer surgery, poor preoperative function was associated with 1-year mortality, and preoperative functional decline and cognitive impairment were associated with 1-year postoperative functional decline.
Surgical decision making for female nursing home residents with breast cancer may need to consider preoperative function and cognitive impairment.
Breast cancer surgery, the most common cancer operation performed in nursing home residents, is viewed as a low-risk surgical intervention. However, outcomes in patients with high functional dependence and limited life expectancy are poorly understood.
To assess the overall survival and functional status changes after breast cancer surgery in female nursing home residents stratified by surgery type.
Design, Setting, and Participants
This study used Medicare claims from 2003 to 2013 to identify 5969 US nursing home residents who underwent inpatient breast cancer surgery. Using the Minimum Data Set Activities of Daily Living (MDS-ADL) summary score, this study examined preoperative and postoperative function and identified patient characteristics associated with 30-day and 1-year mortality and 1-year functional decline after surgery. Cox proportional hazards regression was used to estimate unadjusted and adjusted hazard ratios (HRs) of mortality. Fine-Gray competing risks regression was used to estimate unadjusted and adjusted subhazard ratios (sHRs) of functional decline. Statistical analysis was performed from January 2016 to January 2018.
Main Outcomes and Measures
Functional status and death.
From 2003 to 2013, a total of 5969 female nursing home residents (mean [SD] age, 82  years; 4960 [83.1%] white) underwent breast cancer surgery: 666 (11.2%) underwent lumpectomy, 1642 (27.5%) underwent mastectomy, and 3661 (61.3%) underwent lumpectomy or mastectomy with axillary lymph node dissection (ALND). The 30-day mortality rates were 8% after lumpectomy, 4% after mastectomy, and 2% after ALND. The 1-year mortality rates were 41% after lumpectomy, 30% after mastectomy, and 29% after ALND. Among 1-year survivors, the functional decline rate was 56% to 60%. The mean MDS-ADL score increased (signifying greater dependency) by 3 points for lumpectomy, 4 points for mastectomy, and 5 points for ALND. In multivariate analysis, poor baseline MDS-ADL score (range, 20-28) was associated with a higher 1-year mortality risk (lumpectomy: HR, 1.92 [95% CI, 1.23-3.00], P = .004; mastectomy: HR, 1.80 [95% CI, 1.35-2.39], P < .001; and ALND: HR, 1.77 [95% CI, 1.46-2.15], P < .001). After multivariate adjustment, preoperative decline in MDS-ADL score (lumpectomy: sHR, 1.59 [95% CI, 1.25-2.03], P < .001; mastectomy: sHR, 1.79; [95% CI, 1.52-2.09], P < .001; and ALND: sHR, 1.72 [95% CI, 1.56-1.91], P < .001) and cognitive impairment (lumpectomy: sHR, 1.27 [95% CI, 1.03-1.56], P = .02; mastectomy: sHR, 1.26 [95% CI, 1.09-1.45], P = .002; and ALND: sHR, 1.14 [95% CI, 1.04-1.24], P = .003) were significantly associated with 1-year functional decline across all breast cancer surgery groups.
Conclusions and Relevance
For female nursing home residents who underwent breast cancer surgery, 30-day mortality and survival as well as 1-year mortality and functional decline were high. The 1-year survivors had significant functional decline. This study’s findings suggest that this information should be incorporated into collaborative surgical decision-making processes.
Breast cancer surgery is the most common cancer operation performed in nursing home residents, comprising 61% of such procedures.1 More than half of nursing home physicians report encountering residents with suspected breast cancer through screening or physical examination. A total of 67% of these patients are referred for diagnosis and treatment.2 Because most breast cancer clinical trials have excluded women who are older, frail, and cognitively impaired, treatment outcomes in this population are uncertain.3 Clinical practice guidelines in oncology demonstrate consensus that breast cancer treatment decisions for older women, including surgery, should be individualized based on treatment benefits and harms as well as patient preferences.4 The benefits of breast cancer surgery may not be realized in patients who have competing risks from significant comorbid disease. For those with limited life expectancy, long-term mortality benefits from surgical treatment of breast cancer are unlikely to be gained.
Although operative mortality after breast cancer surgery is low in the general population, the burdens of surgery (eg, functional disability) can be substantial. Regardless of procedure type (eg, lumpectomy, axillary lymph node dissection [ALND]), more than one-third of middle-aged women report poor range of motion or pain that affects their upper body function 6 months after breast surgery, even after lumpectomy alone.5-7 Poor range of motion in the arm can result in diminished ability to perform important activities of daily living (ADLs), such as feeding and toileting independently, in older populations. In nursing home residents with limited life expectancy, because death from other causes may occur before breast cancer causes symptoms and pain, a better understanding of surgical morbidity is important for decision making in this population. In addition, understanding the functional trajectory after breast cancer surgery in nursing home residents is essential for making realistic informed decisions. Using Medicare and national nursing home data, we evaluated mortality and functional decline in female residents who underwent inpatient breast cancer surgery in the United States and identified patient characteristics associated with mortality and functional disability after surgery.
We identified a cohort of female long-term nursing home residents (age ≥67 years) who underwent an inpatient breast cancer surgical procedure by linking data from 100% national Medicare inpatient files (2003-2013) with the Minimum Data Set (MDS) for nursing homes (2003-2013). The Medicare inpatient files contain discharge information for all fee-for-service inpatient hospitalizations for Medicare beneficiaries. The MDS is a mandatory assessment of all nursing home residents who reside in facilities participating in Medicare or Medicaid programs. The MDS assessments of a resident’s health and functional status are administered by nursing staff and are completed quarterly at the time of admission or readmission to the nursing facility and when the resident experiences a change in clinical status. This study was approved by the University of California, San Francisco Committee on Human Research, which determined that informed consent was not required.
The cohort was identified using International Classification of Diseases, Ninth Revision (ICD-9) diagnosis codes for breast cancer and procedure codes for lumpectomy, mastectomy, and ALND with a lumpectomy or mastectomy in the Medicare inpatient file (eAppendix in the Supplement). For the small number of residents who underwent multiple breast procedures, we used the first index case. Patients were classified as long-term nursing home residents if they underwent 2 or more consecutive MDS assessments more than 90 days apart during the 6 months before surgery confirming a preoperative nursing home length of stay for more than 90 days.
We assessed 30-day mortality, 1-year mortality, and 1-year functional decline after surgery. Death was assessed using the Medicare Denominator File. Functional status was measured by trained evaluators using MDS assessments of a resident’s self-performance of the ADLs. Data on performance of ADLs in the MDS include questions about physical function in 7 activities: dressing, eating, toileting, personal hygiene, bed mobility, transferring, and ambulation. Performance of each of these activities is rated using a scoring system from 0 (independence) to 4 (total dependence). The Minimum Data Set Activities of Daily Living (MDS-ADL) score is a sum of these 7 scores and has been validated against standardized measures of function and ranges from 0, indicating independence in all activities, to 28, indicating total dependence in all activities.8 Consistent with previously published studies,9,10 we defined functional decline as an increase of 2 or more points in the MDS-ADL score.
Demographic data were obtained from the Medicare Denominator File and comorbid diagnoses from MDS assessments. Residents were categorized as having cognitive impairment if they had 1 of the following: (1) a dementia diagnosis in the Medicare Provider and Analysis Review (MedPAR) file, (2) a dementia disease diagnosis on their MDS assessment, or (3) measured cognitive impairment using the MDS Cognitive Performance Scale (CPS) or Brief Interview of Mental Status (BIMS). The MDS-CPS (MDS years 2003-2009) and BIMS (MDS years 2010-2013) are validated scales of cognitive function.11 A CPS score of 3 or higher or a BIMS score of 0 to 12 indicates moderate or worse cognitive impairment and has been used to define dementia.12
We defined baseline functional status as the MDS-ADL summary score reported on the most recent MDS assessment before the breast cancer surgery. In addition, we categorized residents as experiencing preoperative functional decline by comparing the 2 most recent MDS reports collected before surgery and determining whether there was an increase of 2 or more points in the MDS-ADL score. To explore the association between preoperative functional status and postoperative mortality and functional decline, we divided nursing home residents into 4 approximately equal-sized quartiles of functional status based on their preoperative MDS-ADL summary score.
For our primary analyses, we examined the association of patient characteristics with outcomes stratified by procedure type. We used the Kaplan-Meier method to estimate cumulative incidence of 1-month and 1-year mortality from the date of surgery. Then, we used Cox proportional hazards regression models to estimate the adjusted and unadjusted hazard ratios (HRs) for 1-year mortality associated with individual resident characteristics. We fit Fine-Gray competing risk models13 to estimate the subhazard ratios (sHRs) for functional decline during the 12-month follow-up associated with individual resident characteristics, taking into account the competing risk of mortality.
In all analyses, the patient was the unit of analysis. Statistical significance was defined as P < .05 using 2-sided tests with a 5% α level. Analysis was performed using SAS statistical software, version 9.4 (SAS Institute Inc). The figures were drawn using R version 3.1.1 (R Foundation for Statistical Computing).
A total of 5969 nursing home residents (mean [SD] age, 82  years; 4960 [83.1%] white) underwent breast cancer surgery. Characteristics of the women in the cohort are presented in the Table. Baseline comorbidities were common: 3396 (56.9%) had cognitive impairment, 2175 (36.4%) had diabetes, 1608 (26.9%) had rheumatologic disease, 1324 (22.2%) had heart failure, and 1315 (22.0%) had cerebrovascular disease. The mean (SD) MDS-ADL score was 13.9 (7.8), indicating a high level of preoperative functional dependence. A substantial proportion of residents (962 [16.1%]) had experienced functional decline in the 3 months before surgery. Breast surgery was performed during an elective admission in 4473 cases (74.9%), and 3661 (61.3%) had an ALND.
The observed 30-day mortality rate was 3.2%, with the highest rate among those who underwent lumpectomy alone (8.5%). In multivariate analysis, for all subcohorts, surgery performed during an urgent or emergency admission was associated with a higher risk of 30-day mortality (lumpectomy: adjusted HR, 4.07 [95% CI, 1.85-8.95], P < .001; mastectomy: adjusted HR, 1.91 [95% CI, 1.09-3.32], P = .02; ALND: adjusted HR, 1.82; 95% CI [1.11-2.99], P = .02) (eTable 1 in the Supplement). For all subcohorts, age, comorbidities, functional decline before surgery, and hospital admission within 1 year before surgery were not significantly associated with 30-day mortality. In the ALND subcohort, severe functional disability at baseline was associated with the highest 30-day mortality risk (adjusted HR, 2.90; 95% CI, 1.32-6.37; P = .008).
One-year all-cause mortality was 31% to 42%. The residents in the most functionally dependent quartiles experienced high mortality in all 3 subcohorts (Figure). In a multivariate analysis, an MDS-ADL score of 20 to 28 (indicating significant ADL dependence) was associated with death at 1 year compared with an MDS-ADL score of 0 to 7 (lumpectomy: HR, 1.92 [95% CI, 1.23-3.00], P = .004; mastectomy: HR, 1.80 [95% CI, 1.35-2.39], P < .001; ALND: HR, 1.77 [95% CI, 1.46-2.15], P < .001) (eTable 2 in the Supplement). In addition, cognitive impairment was associated with 1-year mortality among patients who underwent mastectomy (HR, 1.26; 95% CI, 1.03-1.54; P = .03) and among patients who underwent ALND (HR, 1.22; 95% CI, 1.07-1.39; P = .002) but not among patients who underwent lumpectomy (HR, 1.22; 95% CI, 0.93-1.60; P = .15).
Among patients who survived 1 year, 55% to 60% had a functional decline. These patients experienced clinically significant functional decline at 1 year after surgery: mean MDS-ADL score decreased by 2.8 points after lumpectomy, 4.1 points after mastectomy, and 4.6 points after ALND. In multivariate analysis, for all subcohorts, ADL decline before surgery was associated with an increased risk of 1-year functional decline (lumpectomy: sHR, 1.59 [95% CI, 1.25-2.03], P < .001; mastectomy: sHR, 1.79 [95% CI, 1.52-2.09], P < .001; ALND: sHR, 1.72 [95% CI, 1.56-1.91], P < .001). Cognitive impairment before surgery was also associated with a risk of 1-year functional decline (lumpectomy: sHR, 1.27 [95% CI, 1.03-1.56], P = .02; mastectomy: sHR, 1.26 [95% CI, 1.09-1.45], P = .002; ALND: sHR, 1.14 [95% CI, 1.04-1.24], P = .003) (eTable 3 in the Supplement). With an MDS-ADL score of 0 to 6 as reference, an MDS-ADL score of 20 to 28, which indicates severe functional dependence at baseline, was inversely associated with 1-year functional decline (lumpectomy: sHR, 0.41 [95% CI, 0.30-0.56], P < .001; mastectomy: sHR, 0.49 [95% CI, 0.40-0.60], P < .001; ALND: sHR, 0.54 [95% CI, 0.47-0.62], P < .001).
In a national registry of all elderly women residing in nursing homes in the United States who underwent an inpatient breast cancer operation from 2003 to 2013, more than half were cognitively and functionally impaired before their surgery. Of 5969 breast cancer operations, 3661 (61.3%) were of the most invasive type (ie, a lumpectomy or mastectomy with an ALND). Among 666 patients who underwent the least invasive type of breast surgery (ie, lumpectomy alone), 56 (8.4%) died within 30 days of surgery. Overall, 1842 of 5969 (30.9%) died within a year of their breast cancer surgery, and among those women who survived 1 year, 3478 of 5969 (58.3%) experienced significant functional decline. Poor function before surgery was significantly associated with an increased risk of 1-year mortality across all breast cancer surgery types. In addition, those with the least functional dependency before surgery were most likely to have a 1-year functional decline in any of the breast cancer surgery operations. Cognitive impairment and ADL decline before surgery were independent risk factors associated with 1-year functional decline among patients undergoing any of the breast cancer surgery operations.
Our findings differ from a prior analysis14 that examined mortality after breast cancer surgery. This prior analysis14 of 30-day mortality in community-dwelling women after breast cancer surgery reported a mortality rate of 1%. In the nursing home cohort in our study, 30-day mortality ranged from 2% to 8%, depending on surgery type. The highest mortality rate was associated with the least invasive procedure (ie, lumpectomy, 8.4%), which appeared to be performed in the sickest patients. A higher mortality rate is somewhat expected because of advanced age and increased comorbidities in nursing home residents. However, a 30-day mortality of 8.4% is higher than would be anticipated for a surgical procedure that is generally considered to be low risk. It is likely that the patients identified as high risk were offered lumpectomy only in an attempt to minimize morbidity. In addition, because lumpectomies are typically outpatient procedures, this cohort of women represent patients in whom inpatient stay was deemed to be necessary after lumpectomy alone, likely implying a greater baseline risk than can be captured in administrative databases.
Although this is the first study, to our knowledge, to examine long-term functional loss in older nursing home residents undergoing breast cancer surgery, long-lasting functional loss has been reported in younger cohorts. Regardless of procedure type (eg, lumpectomy, ALND), more than one-third of middle-aged women report poor range of arm movement affecting their upper body function 6 months after breast surgery, with some reporting this problem up to 18 months after breast surgery.5 In the nursing home cohort in our study, 1-year functional decline occurred in 3478 of 5969 patients (58.3%) still alive at 1 year. A higher rate of worsening function after breast cancer surgery is expected in a nursing home population. Because of the high rates of cognitive impairment in this population and the association of cognitive impairment with poor functional recovery after surgical insults among older adults, the likelihood of breast cancer surgery and affecting a frail, elderly woman’s function in the long term is plausible and likely.15
A high proportion of our cohort underwent ALND, which is consistent with a cohort who underwent inpatient surgery. However, because ALND has no proven survival benefit in breast cancer, the potential benefit of ALND would be for staging to determine need for adjuvant therapy, such as chemotherapy, or for locoregional control.16 Given the hesitancy to administer chemotherapy to elderly nursing home residents that was reflected in low chemotherapy use in this population, the staging benefits of ALND were significantly diminished.1 Although omitting surgery for locoregional disease management may be considered to be suboptimal in the treatment of breast cancer, our findings of significant risk of morbidity and mortality after surgical intervention in this patient population may provide evidence that nonsurgical treatment may be preferable.17 Our data do not permit analyzing outcomes by tumor subtype or stage, but multimodal therapy for breast cancer could potentially offer alternative nonsurgical treatment options. Estrogen receptor–positive tumors may be at least temporarily treated with endocrine therapy; for those with bulky adenopathy, radiotherapy without surgery could also be considered.18,19 In addition to the pain reported in prior studies5-7 in younger women undergoing any breast surgery, the consequences of breast cancer surgery in an even older and frailer population are likely to be even greater. For example, a qualitative study by Walter et al20 describes an 89-year-old woman with angina and dementia who was functionally dependent in 2 ADLs and diagnosed with invasive ductal carcinoma. She subsequently received a modified radical mastectomy, with no evidence of cancer in her axillary lymph nodes. The night after surgery, she pulled off all the bandages, requiring restraints, and developed a seroma. She died 15 months later of a myocardial infarction.
The present study provides new benchmark data for women who are long-term nursing home residents and provides essential information for individualizing surgical decision making when breast cancer surgery is being considered. There is a paucity of data about the long-term functional and mortality outcomes of breast cancer surgery in the nursing home population. Realistic, evidence-based information about fatal and functional outcomes is needed to make informed consent and anticipatory guidance meaningful for those who decide to undergo breast cancer surgery.
Although some practitioners, patients, and caregivers believe breast surgery is necessary to prevent morbidity and mortality from breast cancer, the risks of harm may outweigh the benefit in this frail, vulnerable population, in which many have a limited life expectancy. Our study found high mortality rates among nursing home residents who undergo breast cancer surgery. Overall, the nursing home resident mortality rate has been reported to be 25%, and in this surgical cohort, it was 30.9%.21,22 In addition, among those undergoing a more invasive and functionally impairing procedure (ie, ALND), their ability to care for themselves worsened. Breast cancer care should be individualized, goal-oriented care, and hormonal therapy or symptom management should only be considered options for nursing home residents.
Further evaluation of breast cancer surgery in the long-term nursing home population should be performed. First, a study that evaluates specifically nursing home residents undergoing breast cancer surgery in the outpatient setting should be designed. This group will provide information on outcomes after surgery in a likely less frail and sick nursing home population. Second, to better inform patients of outcomes specifically related to the surgery, a study that compares outcomes of nursing home residents with and without the surgical intervention should be developed and studied. Third, decision aids to support informed decision making in a frail nursing home population that is considering breast cancer surgery should be developed and studied. This future direction may further inform surgical decision making in nursing home residents possibly considering breast cancer surgery.
There are potential limitations to consider. First, we did not have information on outpatient services, thereby missing procedures performed in the outpatient setting. However, our goal was to evaluate the sickest patients who would most likely have their procedure performed in the inpatient setting. Second, because the MedPAR file and the MDS do not have information about outpatient services, we did not have information about endocrine therapy, chemotherapy, and radiation therapy that may have been used in conjunction with or instead of inpatient surgery and may have influenced function and mortality. In addition, we did not have information on staging of the breast cancer. Prior research, however, has found that only 6% of nursing home residents with cancer receive radiation therapy or chemotherapy.1 Third, the MedPAR file and the MDS do not contain data about the anesthesia type; therefore, we were unable to assess the association between the type of anesthesia and outcomes. Fourth, we recognize that the patient’s death or worsening function may or may not be directly attributable to surgery. Fifth, our data set cannot identify nursing home residents with newly diagnosed breast cancer who did not undergo surgery or those who underwent outpatient surgery. However, sicker nursing home residents may not be deemed candidates for surgery. Therefore, these data may not be generalizable to the sickest of all nursing home residents and may underestimate the harms of surgery.
Our findings suggest that elderly female nursing home residents who undergo breast cancer surgery are at high risk of 1-year mortality and functional decline. These findings establish benchmark data for this population and provide cautionary information for practitioners considering offering breast cancer surgery with the hope of prolonging life or improving functioning in frail older women with cognitive and significant functional impairment. Surgery often is associated with curing cancer but can worsen other, more life-limiting comorbidities and function of nursing home residents. This information is essential to guide surgical decision making in frail older women with breast cancer and suggests that multimodality therapy be considered as an alternative to surgery in this population.
Accepted for Publication: May 13, 2018.
Corresponding Author: Victoria Tang, MD, MAS, San Francisco Veterans Affairs Medical Center, 4150 Clement St, Room 181(G), San Francisco, CA 94121 (firstname.lastname@example.org).
Published Online: August 29, 2018. doi:10.1001/jamasurg.2018.2736
Author Contributions: Dr Tang had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Tang, Finlayson.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Tang, Mukhtar.
Critical revision of the manuscript for important intellectual content: Tang, Zhao, Boscardin, Sudore, Covinsky, Walter, Esserman, Finlayson.
Statistical analysis: Tang, Zhao, Boscardin.
Obtained funding: Tang, Finlayson.
Administrative, technical, or material support: Tang, Walter, Mukhtar.
Supervision: Tang, Sudore, Finlayson.
Conflict of Interest Disclosures: Dr Finlayson reported being a founding shareholder of Ooney Inc outside the submitted work. No other disclosures were reported.
Funding/Support: This study was supported by grant P30 AG044281 from the National Institute on Aging/University of California, San Francisco Older Americans Independence Center (Dr Finlayson), grant T32 AG000212 from the National Institute on Aging/Research Training in Geriatric Medicine (Dr Tang), and grant KL2 TR001870 from the National Institutes of Health/University of California, San Francisco Clinical and Translation Science Institute (Dr Tang).
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.
Disclaimer: The views expressed in this article are those of the authors and do not necessarily represent those of the National Institute on Aging.
Meeting Presentation: This paper was presented at the following conferences: Scientific Forum of American College of Surgeons Clinical Congress; October 18, 2016; Washington, DC; Annual Meeting of the American Geriatrics Society, May 20, 2017; San Antonio, Texas; Annual Meeting of the Society of General Internal Medicine; April 19, 2017; Washington, DC; 21st International Association of Gerontology and Geriatrics World Congress of Gerontology and Geriatrics; July 24, 2017; San Francisco, California.
Create a personal account or sign in to: