 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1998 The Journal of Bone and Joint Surgery, Inc.
Current Concepts Review - Methods for Locating Missing Patients for the Purpose of Long-Term Clinical Studies*  , LOS ANGELES, CALIFORNIAInvestigation performed at the Department of Orthopaedic Surgery, University of California at Los Angeles, and the Shriners Hospital for Children, Los Angeles
Long-term follow-up studies are essential in the evaluation of the results of medical care. Unanticipated side effects and early promising results that deteriorate with time are not uncommon findings in the evaluation of newly developed medical treatments. In addition, changes in the health-care system are exerting pressure on health-care providers to contain the cost of medical services. In an effort to be more cost-effective, hospitals and physicians are incorporating cost data into the medical decision-making process. The decision to select one treatment instead of another may be made on the basis of the relative costs involved, particularly over the short term. The effect that these economic factors have on the medical outcome may not be known until long-term follow-up studies have been completed. For this reason, outcome studies have become increasingly popular in the effort to determine the overall effectiveness of various medical interventions. Investigators conduct outcome studies to evaluate clinical and functional results, improvement in the quality of life, cost-effectiveness, and other issues in an attempt to determine the relative risk-to-benefit and cost-to-benefit ratios of modern medical treatment. Unfortunately, these longitudinal studies are difficult to perform. Locating and contacting patients after the passage of time often is much more challenging than it first appears. Generally, substantial amounts of time, energy, and money are invested. The longer the follow-up, the greater the investment. The reasons for the difficulty are clear. We live in a mobile society. According to the United States Bureau of the Census, 17 per cent of Americans move each year. Three per cent move to a different county, and another 3 per cent move to a different state9. After ten or twenty years, a sizable percentage of patients will therefore no longer reside at the same address. In addition, patients frequently change their place of employment, choice of insurance carrier, and physician. The average American changes jobs once every three years1. A woman may change her name when she gets married or divorced, and a man may change his name for a variety of reasons as well. Some individuals are incarcerated, and others attempt to become anonymous or go into hiding to avoid the possibility of incarceration. Aging members of the population may be confined to long-term-care facilities, and others die. The list of potential reasons that an individual may be difficult to locate after he or she receives medical care goes on and on. Nevertheless, in order for an investigator to evaluate the effectiveness of a medical treatment over the long term, patients must be located, interviewed, and often examined. The greater the retrieval rate, the more reliable the study. Elimination of patients from a study simply because they cannot be contacted may introduce a reporting bias and alter the findings of the study. Patients who are not found may disproportionately represent those who are dissatisfied and do not want to return to their caretakers. Conversely, patients who have ongoing problems as a result of their treatment may be more likely to return for follow-up. In either case, the assessment of treatment will be biased. Even the loss of a small number of patients can result in changes in the statistical conclusions that may be drawn, particularly if the overall sample size is small and the elimination of patients is anything but random. This bias could prove to be harmful if it leads to the recommendation of one treatment method instead of another on the basis of incomplete or inaccurate data. Ideally, every clinical study would include every patient who received a particular treatment, and reporting bias would be eliminated. In 1971, Chung et al. reported the results of a long-term follow-up study of fifty-six patients who had had a Colonna capsular arthroplasty4. Even though thirty-nine patients had moved and twenty-three had not been seen at the authors' center in fifteen years or more, the authors were able to locate all fifty-six patients an average of seventeen years after the procedure. This success in finding all fifty-six patients was sufficiently unusual that Chung was asked to write a synopsis of the methods that he and his colleagues had used to locate missing patients3. Chung et al. employed a variety of sources, including hospital records, telephone directories, and telephone operators, to gather vital data on each patient. Letters were sent to the last known address of each patient as well as to former neighbors. The United States Postal Service was contacted for change-of-address information, and the records of the Department of Motor Vehicles were searched for vehicle-registration data. Inquiries were made through appropriate professional agencies, and the nursing staff was asked for additional helpful information. Even radiographs were inspected for clues, such as the name of an outside medical facility where the patient might have been known to the staff. Although Chung et al. were able to locate all of their patients4, the entire process took one year3. Not surprisingly, patients who had pain or a poor result frequently did not return questionnaires and tended to be resistant and hostile when they were finally located. All eight patients who were classified as uncooperative or hostile were found on examination to have poor function of the hip. Numerous telephone calls and letters were needed to reach these patients before their follow-up evaluations were completed. Seven of these eight patients could not be located for more than six months. That study3 illustrates two important points. First, locating patients through traditional methods can be effective but requires a considerable investment of time and energy. Second, patients who have had a negative experience or a poor outcome tend to be more difficult to contact. Frequently, these patients will have transferred their medical care to another physician and do not feel obligated to cooperate. Failure to locate and include these patients in a clinical study would alter the accuracy of the conclusions. Although it has been more than twenty years since Chung et al. described their methods for locating missing patients3,4, many investigators still rely on these same techniques. While these methods continue to have their place in the search for missing patients, the introduction of personal computers, the development of large databases, and the widespread availability of the Internet have led to the evolution of more effective techniques for finding missing patients. The purpose of this Current Concepts Review is to familiarize the reader with some of these techniques. Because these methods are dependent on a rapidly evolving technology, some of this information will be outdated shortly. Our intent is to expose the reader to the concept of utilizing electronic databases in the search for missing patients. It is also worth noting that, while some of these methods may seem very deliberate, all are legal and well accepted means of obtaining accurate information.
In the lexicon of investigators who search for missing persons, individuals are categorized into three groups: whereabouts, skips, and criminals. Most missing persons are not actually missing; rather, their whereabouts are unknown to the investigator. Because these individuals, known as whereabouts, are not attempting to conceal themselves from others, they can be readily traced. Skips are individuals who may have failed to pay bills, left families, or defaulted on other obligations. They are not considered criminals, but they do not wish to be found and may be actively avoiding detection. Criminals, in contrast, may go to great extremes in order to hide their location. Unfortunately, there is no way for the investigator to know if a patient is wanted by law-enforcement officials. If the patient is avoiding efforts by law-enforcement officials to find him or her, the scientific investigator's search will be more difficult as well. Whereabouts and most skips can be readily located, whereas criminals are much more difficult to find. Some patients will not be found despite an investigator's best efforts. These patients include criminals as well as victims of foul play, tax evaders, participants in witness-protection programs, and individuals who have falsified their information. An investigator's ability to find all other types of patients is limited only by the amount of time, energy, and money available for the process.
Initiation of the Search Process It is important to note that each child should have a Social Security number by the age of one year for tax-reporting purposes. Therefore, the investigator should ask parents to provide that information as soon as it becomes available. If possible, the names and Social Security numbers of siblings should be included with the information on both parents. Until the child is old enough to work and to drive, he or she will be difficult to find without information about the parents. Similarly, after the parents have died, the patient will be difficult to locate without his or her own unique identifying data.
The United States Postal Service Various alumni associations across the country send out letters each year with a request for change-of-address information from the post office. If the individual has moved within the year, the new address will be returned to the sender. In order for this method to be effective, a stamped postcard must be generated every year for each patient, and time and personnel are required to update the retrieved information. Investigators embarking on a prospective clinical trial could incorporate this method into the protocol.
The Telephone If the investigator is searching for a large number of patients, he or she may prefer to provide the phone company with a list of names and wait for the reply, particularly if there are many area codes in the surrounding region. Some patients may have changed their telephone number to an unlisted number. Unlisted numbers are contained in a separate database that can be searched with the assistance of a supervisor. The rules vary by phone company, but searches of this database generally can be performed if the investigator provides his or her medical license number and explains the medical purposes involved. The databases searched by telephone operators are updated continuously, which is clearly advantageous. White-page listings for the entire United States, however, can be purchased on compact disk (CD-ROM) for a relatively small cost (approximately thirty dollars) and can be searched on any home computer. If the investigator is searching for a large number of patients, this may be a good investment. The information is updated yearly, so it is important to check the package label before purchase to be sure that the disk is current. A source of information that is often overlooked is the crisscross, or reverse, directory. A reverse directory, which can usually be found in the local library, lists telephone numbers by address. This type of directory is not available on compact disk. If an investigator is having difficulty locating a patient's current telephone number because the patient has moved or the number has become unlisted, it is possible to obtain the numbers of the patient's neighbors through this directory. Neighbors can be a valuable source of information. If a neighbor is not familiar with the missing patient, it is possible to ask for the name and telephone number of another neighbor who perhaps has a closer relationship with the patient. If the patient has moved, the neighbor may be able to provide the name of the city where the patient lives. During this process, it is important that the investigator be honest about the reason for the call and be careful not to reveal any protected information. If the patient rented his or her residence, it may be possible to obtain the number of the previous landlord. If a security deposit was owed or if rent was outstanding, the landlord may have already found the patient and be willing to share helpful information. An investigator also may attempt to contact relatives or previous employers of the patient. Nordberg recently described a technique for conducting this type of telephone interview. With permission from the author, we have summarized his technique as follows. Identify yourself by name but do not identify your facility unless asked to do so. Know to whom you are speaking. State the problem and ask for help. Make your request courteous and friendly but sound confident. Encourage the informant through easy conversation by asking leading questions naturally and tactfully. Ask for the new address, place of employment, and telephone number only. Let the person at the other end of the line offer any additional information. Make use of the psychological pause by waiting after each question. Leave it up to the person to make the next move. All questions from the informant should be followed by a combination answer and question. This will keep you in control of the conversation. Know how and when to end your call. Be honest and ethical at all times, and be sure to thank the source for his or her time.
The Internet The drawback of the Internet is that sometimes many sites must be searched before the person in question is found. This can be time-consuming if the investigator is searching for a large number of patients or if the Internet connection is slow. It is important that the investigator try each source before moving on to the next site. Failure to locate a number on one site does not mean that a search of another site will be fruitless. Some of these online databases include: 206.129.166.101 (Infospace "The Ultimate Directory") www.semaphorecorp.com (National Address Browser) www.databaseamerica.com www.four11.com www.yahoo.com/reference/whitepages www.angelfire.com/pages0/ultimates www.yahoo.com/search/people www.switchboard.com www.whowhere.com www.lookupusa.com We have found the angelfire.com site to be particularly useful because it provides access to multiple telephone-number search engines in one location. If a patient is not found on the first attempt, another engine can be searched easily. If the telephone number has been disconnected but an address is given, a personal visit to the patient may be appropriate. The angelfire.com site is capable of generating a detailed map for assistance. The databases differ slightly with regard to the type of information that they will accept for use in the search. Depending on the information available to the investigator, one site may be more helpful than another. With some effort, these sites can also be used to find the telephone numbers of neighbors who live on the same street where an individual used to reside; this method is similar to the technique of using a crisscross, or reverse, directory.
The Department of Motor Vehicles
Voter Registration Files
Credit Bureaus Credit bureaus acquire information from a number of different sources, including all types of financial statements, motor-vehicle registrations, magazine subscription lists, voter registration files, white-page listings, postal address updates, the National Death Index, and so on. Their databases are immense and are growing at an astounding pace. Most financial institutions and hospitals already have access to at least one of these services for the purpose of verifying credit information. Access is easily obtained though a modem connection with use of customized software. Consumer inquiries can be performed quickly and effectively. The Fair Credit Reporting Act10 stipulates that credit information may be requested from a consumer reporting agency for certain permissible purposes only. Obtaining a credit history without cause is a federal crime and can lead to serious penalties, including fines and imprisonment. However, it is not necessary to obtain a complete credit history in order to find a patient. As the Fair Credit Reporting Act does not contain an exact definition of what constitutes a legitimate business need and does not specify the procedure for verifying such a need, personal information held by credit agencies is readily available to anyone who is willing to pay for it. Each of the major credit-reporting services has developed products that are ideal for the purpose of locating missing patients. For a little more than one dollar, the vital information on any individual can be obtained without a formal credit check. This avoids any legal problems and greatly reduces the cost of the search. Vital information includes the patient's name, current and previous addresses, Social Security number, age, employer, and spouse's name. A positive search result should provide the investigator with enough information to contact the patient by letter or telephone. TRW offers two products, TRW Social Search and TRW Address Update, which can be used to search a database that contains information on more than 190 million individuals. TRW Social Search provides the individual's name, year of birth, current address and as many as ten previous addresses, spouse's name, and employer after the individual's Social Security number is entered. As many as twenty Social Security numbers may be entered at once for efficiency. If the Social Security number is not known, the patient's name and last known address are sufficient to perform a search with use of TRW Address Update. The output includes the patient's current address and Social Security number. TRW does not provide current telephone numbers; however, if desired, a current telephone number should be easy to obtain with use of the patient's current address, as described previously. Transunion offers four different products, which can be used to search two separate databases. TRACE, TRACEplus, and ReTRACE all work from the same database, which contains records on more than 220 million individuals. TRACE is an automated national search system that enables the investigator to ascertain or confirm a patient's name and address by entering the patient's Social Security number. TRACEplus is a more informative search system that provides the patient's name, address, aliases, date of birth, telephone number, and employment information (when available). ReTRACE provides the individual's Social Security number, telephone number, and date of birth, and as many as three addresses, after the individual's name and last known address are entered. Transunion also maintains another database, the ATLAS directory, which is compiled from a number of different sources, including telephone directories, marketing companies, and the circulation departments of popular magazines. A variety of different types of reports can be generated from this database depending on the data that are available for input and the desired output. An investigator can use the ATLAS database to generate a list of the patient's current and previous neighbors and to perform surname searches according to geographic area. The ATLAS directory currently contains records on more than 150 million individuals. Equifax offers two products, DTEC and ID REPORT, that can be used to find missing persons, again without violating the Fair Credit Reporting Act. With use of DTEC, the investigator can search the nationwide Equifax database of 190 million consumer files by entering the individual's Social Security number. The output includes the patient's name and telephone number as well as his or her current and former addresses. With use of ID REPORT, the investigator is able to search through the same database by entering the patient's name and last known address; the output includes the most current reported information. Overall, these databases are the most comprehensive tracking systems available. Gaining access to these sources is easy and completely legal. Although an internal electronic record is made of every inquiry, no external mark appears on the patient's credit report. Systematic updating makes these databases extremely accurate. These systems provide instant access to hundreds of millions of individuals for a relatively minor cost; most searches cost less than a dollar and a half each. With the exception of the ATLAS directory, which charges ten cents, none of these systems charge for a search that fails to yield data. All of these databases are remarkably efficient in terms of both time and cost. The necessary software can be easily installed on a personal computer that is equipped with a modem. A start-up cost is involved, and monthly charges apply whether the investigator uses the system or not. Most hospitals have the systems running and can request user-specific codes for accurate distribution of billing.
Letter-Forwarding In order for a letter to be forwarded, certain criteria must be met11. First, the investigator must have a strongly compelling reason for wanting to locate the missing person. Second, the investigator must demonstrate that the missing person would want to be aware of the contents of the letter. Third, all other methods of contact must have been exhausted. To have a letter forwarded, the investigator must send a cover letter to a local Social Security office, stating the reason for needing to contact the missing person. The letter to be forwarded must be submitted in a plain, unsealed, unstamped envelope bearing only the missing person's name and Social Security number. A doctor or other health-care professional who wishes to contact a person for the purpose of health-related follow-up will be charged a fee (approximately three dollars). The fee is waived if the investigator is attempting to inform the person of a disease about which the person might be unaware. Inquiries involving 100 individuals or more should be sent to OEER, 6401 Security Boulevard, Baltimore, Maryland 21235. OEER will respond directly to such inquiries and provide additional instructions. The investigator should be aware that many employees in the Social Security Administration are unaware of the written policies regarding the forwarding of letters. If the investigator is polite and patient and emphasizes the importance of the request, his or her effort may be fruitful. If difficulties are encountered at the local Social Security Administration office, it is possible for the investigator to appeal to the main offices in Baltimore. The amount of time that elapses between the date on which the request is made and the date on which the letter is forwarded generally is about two months.
Other Sources FINCEN is the federal government's Financial Central Record Database, which contains records of financial transactions involving all Americans. The database includes information on credit-card transactions, bank accounts, taxes, and so on. This source, which has been called the granddaddy of all databases, is accessible to authorized governmental employees only8. Although the scientific investigator cannot use FINCEN to locate patients, the size of this database illustrates that massive amounts of information pertaining to the nation's residents are continuously being collected and recorded in computerized format. National Credit Information is a private entity that maintains another enormous database. Unlike FINCEN, however, this database is available to the public. National Credit Information provides information to attorneys, professional brokers, private investigators, collection agencies, credit services, property managers, and employers. This one service provides access to vast amounts of information, including consumer credit information (350 million files), Social Security number-tracing (300 million files), address-updating (250 million files), nationwide crisscross telephone and address information (ninety million files), nationwide driving information, nationwide alpha name searches, nationwide Workers' Compensation information, nationwide criminal history records, and death record searches through the Social Security Master Death Record File (forty-one million files). Not everyone qualifies to use this service. The user must be associated with a legitimate business and must have a legitimate need for access. There is a substantial start-up fee, and charges (usually between five and fifty dollars) are incurred for each report generated. Although this database is extremely powerful, it is also expensive. Fortunately, most of this same information can be obtained from other sources for far less money.
In some cases, an investigator will be unable to reach a patient because the individual has died. In most instances, this information is contained in the patient's credit history and can be found with the use of one of the aforementioned credit-reporting services. Other sources that can be used to determine if a specific patient is still alive include the Social Security Administration, the Internal Revenue Service, the Veterans Administration, and the National Death Index. In a study comparing the effectiveness of these sources2, the National Death Index proved superior. Specifically, the National Death Index accurately reported 97 per cent (155) of 160 deaths; the Social Security Administration, 83 per cent (337) of 407 deaths; the Veterans Administration, 80 per cent (326) of 407 deaths; and the Internal Revenue Service, only 23 per cent (ninety-two) of 407 deaths. The National Death Index, instituted in 1979 by the National Center for Health Statistics, is a central computerized index of death-certificate information obtained from each state's vital records office. It is updated twelve months after the end of each year. Researchers must submit a nine-page application before the initial search; subsequent searches require a quick approval. The cost, based on usage, is approximately ten cents per person per year searched7. The information from the National Death Index is entered into the databases maintained by the three major credit bureaus soon after it is updated annually. Although the records of the Social Security Administration are confidential and are not available for review by the public or even by law-enforcement officials, the office will generally verify whether a death claim is on file. The Internet can also be helpful and efficient. The Death Master File from the Social Security Index (version 3Q95) contains fifty-one million records for review and is easily accessed at the Web site www.ancestry.com. Although the records of the Social Security Administration were found to be only 83 per cent effective overall for identifying dead individuals2, they appear to be much more effective for locating dead elderly patients; success rates of as high as 97 per cent (155 of 160) have been reported when these records were used to search for patients who were more than sixty years old at the time of death5,12. Additional sources that can be consulted to determine if a patient has died include the Deceased Individuals File and the Equifax Nationwide Death Search. The Deceased Individuals File, which is accessed by entering "go computrace" after connecting to Compuserve, contains information on United States citizens who died after 1928 and for whom a death claim was filed with the Social Security Administration. This database is much easier to use when the investigator knows the patient's date of birth and Social Security number. The Equifax Nationwide Death Search allows the investigator to search sixty-one million records from more than 100 sources. If the investigator has already obtained the header information from the credit report through Equifax, this search will be redundant. Rich-Edwards et al., in a study of the ability of researchers to determine the vital status of individuals, reported that this method had a sensitivity of 79.2 per cent (156 of 197) and a specificity of 99.9 per cent (1996 of 1997).
Once the investigator has diligently applied the methods described here, most missing patients will have been located or found to have died. At that point, the investigator must decide how important it is to contact the remaining patients. If necessary, the investigator may seek the assistance of an information broker. Information brokers are specialists who gather and sell information regarding individuals. They have access to literally hundreds of additional sources of information, including numerous computerized databases that have not been mentioned, all of which can be used to locate missing patients. These databases are extensive and vary in terms of both content and method of data acquisition. Information brokers can use these sources to rapidly check public records of marriage and divorce proceedings, real estate holdings and contracts, bankruptcies, liens, foreclosures, declarations of homestead, probate proceedings, building permits, tax liens, professional licenses, partnerships and corporations, fictitious name filings, driving records, motor-vehicle registrations, and civil and criminal proceedings. Most of these additional databases are restricted to the general public and require a substantial annual membership fee as well as usage fees. Because of the extensive resources available to professional information brokers, a comprehensive listing of their methods and available databases is not possible. The investigator should be aware that the use of an information broker can be very costly; however, volume discounts may be negotiated in certain cases. Another expensive option is to hire a private investigator. This option may be useful if the person to be located is known to be residing in a certain part of the country that the scientific investigator is unable to access personally.
The need for long-term follow-up studies in the evaluation of medical care is not unique to the field of orthopaedic surgery. Medical treatment is continuously changing, and there is pressure from many sources to make it more effective and cost-efficient. The impact that these factors will have on the lives of patients cannot be determined without investigative outcome studies. In order for those data to be meaningful, every reasonable effort should be made to obtain information on all patients in a study. Longitudinal studies require continued contact and evaluation of patients for many years after the administration of treatment. When planning a clinical study, the investigator should include the expense of locating patients in the estimates of the study costs, regardless of whether the study is prospective.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
  CiteULike  Connotea  Del.icio.us  Technorati What's this?
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Introduction
Sources for Locating Missing...
1681 et seq. Congressional Record. Washington, D.C., 1970.
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |