|
As clinical trials become more globalized, drug development companies must work around communications challenges including language barriers that have widespread implications for data collection and analysis. This article focuses on the challenges presented in the deployment of technology solutions used to improve the management of studies and study data collection. We consider three main areas: centralized randomization and trial supply management using Interactive Voice Response (IVR) systems, collection of electronic Patient Reported Outcomes (ePRO) data, and Electronic Data Capture (EDC) systems used in the collection of site clinical assessment data.
Bill Byrom, product strategy director, ClinPhone, Nottingham, UK;
Andrew Scobbie, telecoms manager,
ClinPhone, Nottingham, UK;
Maureen Smith, linguistics manager,
ClinPhone, Nottingham, UK
& Dennis Wijnker, principal architect, EDC, ClinPhone, Irvine, CA, USA.
Introduction
Over the past two decades, clinical trials have become larger and more complex and this trend is continuing. The U.S. Office of Inspector General reports that an average of 4,237 subjects were used in New Drug Applications (NDAs) to the FDA from 1994 to 1995 compared with 1,321 subjects from 1981 to 1984[i]. Inclusion criteria in many studies have become more specific and restrictive, while the number of clinical trials actively seeking patients has increased. These factors affect the ability of sponsors to recruit patients in a timely manner. Moreover, the cost of clinical research continues to rise.
All these factors have influenced the current globalization of clinical trials. The major locations for drug development remain North America, Western Europe, and Japan, but markets such as those in Eastern Europe, Asia, Latin America, and Africa have emerged as important locations for clinical research. These countries have a sizable population of patients, and can often provide a source of drug naïve patients — a growing requirement in some therapy areas. Moreover, the infrastructure and quality of clinical research conducted in many of these counties is of a very high standard.
For example, Synexus, a patient recruitment and site management company, has in the last 12 months expanded from solely U.K. located study sites into these growing regions by opening centers in Poland, Bulgaria, Hungary, India, and South Africa — the rationale being patient access for international clinical trials. These trends are confirmed by the 2001 report from the U.S. Office of Inspector General[ii], which states that over the ten years from 1990 to 1999, NDAs containing clinical research conducted in new countries has risen rapidly — from 28 to 79 countries, with the largest growth being seen in Eastern Europe, Latin America, and Russia (Table 1). Sponsors also benefit from conducting clinical trials in the expanding list of potential important markets for the approved product. This globalization brings with it a multitude of challenges.
Speaking the Right Language
IVR solutions are often used in large, multinational clinical trials to simplify the logistics of managing site and depot medication stock and in providing robust and reliable randomization[iii]. The multilingual nature of clinical trials, however, presents some unique challenges in the effective use of these solutions. At ClinPhone, a global clinical technology provider, we are well aware of the growing internationalization of trials. In November 2006, for example, the top 20 countries accessing our IVR systems included Poland, Hungary, Russia, Ukraine, Argentina, Brazil, Mexico and South Africa (Table 2).
Although clinical study sites often use English documentation, such as the study protocol and case report forms, when interacting with IVR it is common for communication with site staff to be achieved using their native language.
The telephone numbers used by study sites to access an IVR system are different for each country. This means that when an incoming call is received, an IVR system can determine the country of origin and play an initial welcome message in the appropriate language. Because many countries are multilingual, language requirements can also be determined at the user level — upon entry of a user code, the system is able to determine the individual language requirements of the user and make a switch if appropriate. For example, a French-speaking site user calling from Switzerland may be initially greeted in Swiss German, but upon entry of their unique code and PIN will hear remaining dialogue in French.
The translation and construction of IVR prompts (recorded messages) is more complex. Ideally at each step in a call, an IVR system should play a specific prompt to guide or relay information to the user. Delivering this in an efficient way imposes a specific structure on the way sentences should be constructed. For example, when confirming the entry of a patient’s screening number the system might report: “You entered 104”. To achieve this, the message itself is constructed be playing a pre-recorded prompt “You entered” and reading the data currently stored in the system database: “1”, “0”, “4”.
In English, this sentence construction is in common with the normal use of the language: Subject, Verb, Object (SVO). However, in other languages sentences are normally constructed using different word order. In Korean and Japanese, for example, it would be more natural to say, “You 104 entered” (SOV). The sentence structure imposed by the construction of prompts, in this case reading the screening number from the database at the end of the sentence, creates a challenge for those providing translations. Experienced translators can provide optimal sentence construction to serve the set of languages required by the study, and construct sentences using that structure that do not sound out of place in each language.
Translation is not the only challenge. Although text-to-speech engines are improving, they have yet to catch up with the number of countries required in today’s clinical trials and can only represent a restricted set of languages adequately. Local language speakers are optimal in providing quality recordings of the study prompts, and for some languages or countries these can be difficult to source.
Providing Robust Communications
Most IVR vendors provide toll-free access for sites when making IVR or support calls. Toll-free numbers offered for calls terminating within the same country are straightforward, as the cost can be charged by the telecommunications provider to a particular business or organization. However, in international clinical trials, calls from around the globe need to terminate in a single location to enable central control of the study.
International toll-free numbers rely on an agreement between telecommunications providers to exchange traffic. For example, a South African telecommunications provider (e.g. Telkom SA Ltd) may allow a U.K. provider (e.g. British Telecom) to sell toll free numbers in South Africa which terminate in the United Kingdom, on the basis that they would be able to supply toll-free numbers that terminate in South Africa. In each case, the provider only makes revenue from the calls terminating in their territory, and so there has to be a mutual business interest in pursuing this reciprocal agreement.
Because such an agreement could be rather one-sided, telecommunications providers are not always able or willing to establish toll-free capabilities in all countries. For this reason, to provide a robust and reliable service, IVR vendors seek to obtain toll-free capabilities on more than one network for each country and additionally have international direct-dial numbers as an emergency fail-safe.
In countries where toll-free service cannot be arranged, calling cards can enable sites to make free calls via a two-stage dialing approach. These solutions have enabled IVR providers to offer practical, centralized trial management solutions across the globe.
Additional Language Considerations: ePRO
When communicating directly with patients using IVR, particularly when collecting patient reported outcomes, authentic local language voice recordings are essential to ensure patient understanding and acceptance. In addition, it is sometimes appropriate to have a specific gender voice to deliver the questions and dialogue — for example, when delivering a diary for therapy areas such as sexual dysfunction.
However, when collecting patient self-report data — whether as a primary or secondary endpoint — it is of vital importance that you are “speaking the same language” as the patient. Usually patient reported outcomes instruments are developed in English. To be used in a multinational clinical trial, however, researchers must be confident that an instrument’s questions are understood in the same way and address the same fundamental concepts in all language groups involved in the study. This will ensure that data collected are independent of country, language or culture, and that the data reported by patients within each country can be pooled in the overall study analysis.
The process of demonstrating that a patient reported outcomes instrument is equivalent across countries involves linguistic validation. This may be achieved following a process of forwards and backwards translation followed by review by a small local-language medical panel to validate the meaning of a questionnaire and verify that the interpretation of each item is equivalent to that intended in the original language version. In some cases, however, the process may include cognitive debriefing studies among the target population.
Interactive Voice Response systems provide a simple and intuitive way to collect patient reported outcomes data in clinical trials[iv]. Like other electronic methods, they provide the ability to improve data quality using in-built logic- and error-checks to prevent missing, conflicting, ambiguous, or out-of-range data. In addition, time-and-date stamping data entries combined with defined time windows within which entries must be made, improves the contemporaneous nature and integrity of data collected.
However, the advantage of IVR over other means of electronic collection of patient reported data, including using hand-held devices, lies in the ubiquitous nature of the telephone. Most people own or have access to a phone, and are familiar with its use; it is not normally necessary to supply and support hardware for the patient. Moreover, due to the range in functional literacy observed across the populations of even highly developed countries, the spoken word may be more reliably comprehended than the written word. When it is vital to “speak the same language” as the patient, this may be an important consideration.
Challenges in Multilingual Electronic Data Capture
Electronic Data Capture (EDC) systems promise to improve the quality and speed of data collected during site-based clinical assessments by replacing the paper Case Report Form (CRF) process with web-forms. In many clinical trials, CRFs are produced in English for all study sites, independent of location. However, this is not acceptable in some countries and many EDC solutions are evolving to make electronic CRFs available in other languages.
In particular, due to the increasing volume of clinical development performed in Japan, and the size of the pharmaceutical market in that country, Japanese (Kanji) is perhaps the highest priority in terms of provision of EDC capabilities in a language other than English. Unlike English, however, Kanji contains over 20,000 unique characters — a challenge in itself.
Rather than enabling an EDC solution to simply operate in another language, it may be desirable to facilitate operation in multiple languages simultaneously. For example, an application deployed in Japan using eCRFs written in Kanji may need to be reviewed in English by the central sponsor personnel. In this scenario, it may be possible for Sponsor staff to view the pre-defined fields and labels that make up the eCRF in English, while these labels and fields are displayed in a second language to the site end-user.
This may require the EDC solution to create the eCRF dynamically by reading the language-specific fields and labels from a central database. As much of the data collected on an eCRF is either numeric or selected from a list, this facilitates the ability to switch the view of the data between languages. However, certain end-user data inputs such as the contents of free-text fields, data comments and responses to queries cannot be converted as the ability to reliably translate “on-the-fly” is impractical.
Much of this capability is enabled by the storage of characters using the Unicode standard, which provides a unique numeric reference for every character independent of the language. In the future, like IVR, this may provide sponsors with the ability to deploy EDC in all the languages used by end-user site personnel.
Conclusions
Clinical trials continue to become increasingly cosmopolitan. To maintain high quality clinical research, sponsors must ensure they are not speaking a foreign language. Technologies such as IVR and EDC provide the means to streamline clinical development, and are key to communicating clearly in this multilingual environment.
References
i U.S. Department of Health and Human Services, Office of Inspector General, Recruiting Human Subjects - pressures on industry sponsored clinical research. June 2000 (OEI-01-97-00195).
ii U.S. Department of Health and Human Services, Office of Inspector General, The globalization of clinical trials: a growing challenge in protecting human subjects. Sept 2001 (OEI-01-00190)
iii Byrom B. Managing the medication supply chain process using clinical technology solutions. GOR, Feb 2005.
iv Byrom B. Electronic diary solutions: Enhanced collection of patient reported outcomes data. EBR, 2004.
|
