The CARE-FALLS Study Protocol for assessing the accuracy of 'stops walking when talking' in the diagnosis of falls
a multinational e-mail/web-based study
Copyright ©1999-2002, CARE-FALLS
This study is funded by a grant from the Physicians' Services Incorporated Foundation.
* Completed October 2000 *

1. We're planning to begin this study at the end of October and hope that you are interested in joining us. Please review the protocol and forward to us any comments or criticisims at carestudy@rogers.com by October 15, 1999.
2. We will recruit patients from November 1, 1999 to March 31, 2000. Following their initial clinical assessment, patients will be followed up for 6 months and data collation and analysis will occur once the Co-ordinating Centre has received all data.
3. Each participating investigator should take responsibility for finding out and adhering to their local policy for the ethics approval of this study.

The elderly population is increasing exponentially and clinicians must seek to optimise the care of these individuals. Falls are the leading cause of accidental death among people aged 75 years of age and older and are also responsible for significant morbidity including fracture, impaired mobility, fear of falling and admission to long-term care facilities.^[1-5] Several risk indexes for predicting falls have been developed but unfortunately, busy clinicians have little time to complete complex assessments and there is a need to develop and validate screening tools that are quick, inexpensive, and easy to administer. If those patients who are at high risk for falling could be easily identified, clinicians could then do a more complete assessment of these patients and target them for interventions that are known to reduce falls and injuries from falls (such as prescribing hip protectors which have been shown to reduce hip fractures in frail older people living in residential care^[6]).

In a small study, Lundin-Olsson and colleagues observed that patients living in a long-term care facility who were observed to stop walking when a conversation was initiated, were more likely to fall over the next 6 months (likelihood ratio 10) than those patients who were able to continue walking and talking.^[7] This test could potentially be performed by family physicians, hospital-based physicians, nursing staff and physiotherapists amongst other health care professionals. Moreover, this simple manoeuvre, if validated in larger studies in a variety of settings, could allow clinicians to identify the people who would benefit most from interventions proven to decrease the risk of injury due to falls rather than offering the intervention to everyone.

The objective of this proposed study is to determine amongst elderly, ambulatory, talking individuals (living independently or in long-term care facilities), in a wide variety of settings and from primary or secondary care, what is the accuracy of 'stops walking when talking' for predicting falls in the next 6 months?

1. Kannus P, Parkkari J, Koskinen S, et al. Fall-induced injuries and deaths among older adults. JAMA 1999;281:1895-9.
2. Tinetti ME, Williams CS. Falls, injuries due to falls and the risk of admission to a nursing home. N Engl J Med 1997;337:1279-84.
3. Berg WP, Alessio HM, Mills EM, Tong C. Circumstances and consequences of falls in independent community-dwelling older adults. Age Ageing 1997;26:261-8.
4. Rizzo JA, Friedkin R, Williams CS, Nabors J, Acampora D, Tinetti ME. Health care utilisation and costs in a Medicare population by fall status. Med Care 1998;36:1174-88.
5. NHS Centre for Reviews and Dissemination. Preventing falls and subsequent injury in older people. Eff Health Care 1996;2(4):1-16.
6. Lauritzen JB, Petersen MM, Lund B. Effect of external hip protectors on hip fractures. Lancet 1993;341:11-3.
7. Lundin-Olsson L, Nyberg L, Gustafson Y. "Stops walking when talking" as a predictor of falls in elderly people. Lancet 1997;349:617.

Among consecutive elderly ambulatory, talking individuals (both living independently and in long-term care institutions) what is the accuracy of 'stops walking when talking' for predicting falls in the subsequent 6 months?

1. Investigators will join the study in teams of 2 or more people to enable the independent blind comparison of the clinical findings with the occurrence of a fall (one investigator will perform the clinical assessment and the other will provide the follow-up and will be blind to the results of the clinical examination).
2. Investigators will be recruited from centres around the world via the Internet and via the evidence-based health care discussion group.
3. Investigators will include clinicians from both primary and secondary care settings:
1. Primary care settings - includes first contact clinicians seeing self-referred patients. Examples include Commonwealth GPs, Family practitioners, US-style General Internists, some Emergency physicians and other clinicians who provide first contact care.
2. Secondary care settings - includes second contact clinicians who see patients referred by other clinicians. Examples include General physicians, geriatricians and some Emergency physicians who see patients referred to them by other clinicians.
4. Investigators will enrol at least 12 patients per participating team including: 4 patients with a history of falls; 4 patients suspected to be at risk for falls; and 4 patients without a history of falls and who aren't suspected to be at risk for falls.
5. Investigators will use email and the Internet for enrolling in the study, data submission, presentation and discussion of study results and for the generation of manuscripts.
6. Each investigative team will take responsibility for obtaining local ethics approval for the study and proof of this must be forwarded on request to the study co-ordinating centre in Toronto.

1. Patients will be admitted consecutively (i.e. consecutive eligible patients that the investigators sees in a time period specified by the investigator). For example, patients can be recruited on different clinic days but would still be considered consecutive, eligible patients.
2. Patients will be of 3 sorts:
1. Those with a history of >= 1 fall in the past 5 years;
2. Those with no previous falls but whom the clinician suspects may be at high risk (uses cane, uses walker/Zimmer frame, has impaired mobility, has poor balance, uses psychoactive medications, or has visual impairment); and
3. Those with no previous falls and not suspected by the clinician to be at high risk of falls.
3. Exclusion criteria include:
• Age less than 75
• Patients using wheelchairs
• Patients with life expectancy < 6 months
• Patients who are unable to speak
• Patients who are legally blind
• Patients who refuse to participate
• Patients with significant cognitive impairment ( Short Portable Mental Status Questionnaire score of <6)
• Patients who have significant hearing impairment (e.g. unable to hear their name being called in a normal speaking voice)
4. Patients meeting the eligibility criteria will be approached for consent to join the study and will be informed that during the subsequent 6 months, an investigator will contact them by telephone on a monthly basis. Patients will be blind to the purpose of the study.

1. Information will be collected on the following items:
1. Age
2. Sex
3. Living arrangements
4. Previous history of falls
5. Risk factors for falls
6. Score on Short Portable Mental Status Questionnaire
2. 'stops walking when talking' - Accompanied by an investigator, each patient will be asked to walk a distance of 10 metres. As the patient begins walking, the investigator will pose the question, 'Tell me what you had for dinner last night.' The investigator will then note whether the patient stops walking for a period of 3 seconds or longer when talking which is considered a positive test result. The investigator should begin by matching the patient's stride and should not slow down or stop unless the patient does. Clinicians should not stop unless the clinician gets >= 3 strides ahead of the patient.
3. The time taken to do the 'stops walking when talking' manoeuvre will be measured.

1. The primary outcome will be the number of people who experience a fall within the 6 months following the clinical assessment.
2. Falls are defined as: an incident in which a patient suddenly and involuntarily comes to rest on the ground or a surface lower than their original station and is not caused by external force such as MVC.
3. For patients living in the community, each month after the clinical assessment, a second investigator (blind to the result of the test) will telephone the patients or caregivers to determine if the patients experienced a fall in the preceding month. For patients living in long-term care facilities, the investigator will contact patients and review the medical records to determine if the patients experienced a fall.
4. Additional outcomes that will be recorded include: the number of falls, the number of falls resulting in medical care; the number of falls resulting in injury and the number of falls resulting in fracture.

Data submission will be by an internet data collection system. Patient identity will be anonymous. Each investigator should have their own method of assigning patient ID numbers which should also be entered on the data collection sheets. To qualify for the study, each patient should have both a clinical exam data collection sheet and a falls data collection sheet. Data collection forms can be downloaded as an Adobe PDF file by clicking here.

For investigators who haven't registered for FALLS yet:

1. Register for the study using the online form at http://www.CAREStudy.com/FALLS/Registration.asp (or click on 'FALLS' and 'Go' in the Data Entry drop down box on the CARE home page.  This takes you to a login page and prompts you to click if you haven't registered as an investigator yet.)
2. Once you have completed and submitted this registration form, you will be issued with an Investigator ID and asked to supply a password for subsequent data entry.  
3. Once we confirm your information, we'll add you to the lists of both CARE members and FALLS investigators.
4. Once you submit your password, you will be welcomed to FALLS.  To enter data on a patient, click on 'FALLS' and 'Go' in the Data Entry drop down box.
5. Enter your patient's ID number and complete the data entry form. Investigators should use their own system of numbering patients. You'll notice at the top of the page there will be a list of any patient ID numbers that you have already used.
6. Once you have entered all data, click "submit". The data will then be reformatted and you will be asked to check the data you entered. If you have made any errors in data entry, you can make them on the previous page. If you want to revise any data, click 'No' at the bottom of the screen in response to the Submit query.Once you are satisfied with your data, click 'Submit'. If you have used a patient ID number which you have used previously, you will be asked to assign a new patient ID number and to resubmit the data. If you leave any information incomplete, the data won't be accepted and you will be prompted to supply the missing data. If after completing this process, you recognise an error has been made in data entry, contact us at carestudy@rogers.com
7. To enter data on another patient, click 'FALLS' and 'GO' in the Data Entry drop down box.

For investigators who have already registered and received an ID number and password:

1. Go to the login page at http://www.CAREStudy.com/FALLS/Login.asp  and enter your ID number and password.
2. Follow steps 5 to 7 above.

The sensitivity, specificity and likelihood ratio for 'stops walking when talking' will be calculated using the number of patients who fall at 6 months as the reference standard. The same calculations will be performed using the number of patients who fall at 3 months and using other references standards including the number of falls and the number of falls resulting in injury.

Patients will be enrolled from November 1 to March 31, 2000. Following their clinical assessment, they will be followed up for 6 months and data collation and analysis will occur once the Co-ordinating Centre has received all data. Considering the start-up period and the period of data analysis, the study will require approximately 9 months to be completed.

1. First analyses and draft reports will only be sent to participating investigators.
2. After the investigators are satisfied with the analyses, and manuscript, it will be submitted by mutual agreement with authorship that includes the phrase 'on behalf of the CARE-FALLS Study Group'.
3. Publications will include a list of all participants in descending order, who submitted usable data on 12 or more patients and all methodologists involved with the study.

* Any comments or concerns about this protocol, please email carestudy@rogers.com.