Clinical Review

CASE A medical practice in disarray

An ObGyn reported the following signs of a problem to a colleague: “Our practice was literally drowning in paperwork. An exam room was recently converted to hold more charts, and 2 warehouses held our overflow. Employees were constantly searching for records, and telephone messages were delayed for hours or days until the chart could be reviewed. Notoriously bad handwriting and incomplete documentation hampered good communication and good medical care. Transcription costs were out of control. Forms helped but added to the ongoing costs and storage problems.”

What are the treatment options?

Electronic medical records (EMR) have progressed from arcane, slow, cumbersome documentation systems to sophisticated, complex, comprehensive ones. These modern systems hold the potential to reduce administrative and management costs by 30% or more, improve clinical workflow, reduce medical errors, facilitate communication between patient and physician, and enable analysis of data for best practice methods, best outcomes and identifying risks and complications.

For practices like the one described in the preceding paragraph—not a fictional account but actual testimony provided by an ObGyn—EMR offer a powerful potential solution to the problems that result from an overwhelming amount of paper documentation, correspondence, charting, claims, and financial transactions. In this article, I offer a general introduction to EMR; in the next (August) issue of OBG Management, I’ll speak with a group of ObGyns and medical practice managers about their experiences—and inexperience—with EMR.

Progress and paradox

Physicians and scientists have made substantial progress over the past 25 years in pharmacotherapeutics, diagnostic technology, procedures, and treatment protocols. In obstetrics and gynecology alone, consider the array of technologies—3-dimensional ultrasonography, minimally invasive surgery, receptor-specific drugs, in vitro fertilization, long-acting reversible contraceptives—that have advanced the quality and effectiveness of care. Yet little progress has been made in the process of caring for patients.

The fact is that physicians, and other health-care providers, are rooted in paper-based processes that sustain inefficiencies, increase costs, and defy the gains that other industries have made by adopting electronic technologies for handling information. Why are we so stuck?

The state of EMR

EMR—of varying functionality—have been available for longer than 20 years. Early models were developed by physicians who had an interest in software coding and design, and were of limited functionality, arcane, and difficult to use in a clinical setting. Some of those early models, and even a few commercial systems in use today, rely on scanning paper documents into computer files. Such systems may eliminate some paper and facilitate document retrieval, but they do nothing to ease management of the complex transactions of health care, and they do not address handwriting illegibility.

Development of complex EMR systems was limited by primitive technology, inadequate distribution channels, and programming that was cumbersome and expensive to maintain. But these barriers have been overcome with fast processors, inexpensive and abundant memory, broadband Internet connectivity, and programming languages that facilitate automated software development.

Modern EMR systems are not simply data repositories: They also support workflow from the beginning to the end of a patient’s consultation with a health-care provider—an event that generates multiple transactions with multiple recipients. A single consultation may, for example, generate orders for lab tests, imaging studies, a surgical procedure, consultation with other physicians, prescriptions, and counseling, and record the subsequent financial transaction. EMR systems by necessity interact with multiple organizations, institutions, instruments, and other software systems. To software developers, and to the clinicians who use their systems, the challenge is to deftly navigate the complexities of health care.

Forces accelerating adoption

Momentum from the Executive Office. In 2004, President George W. Bush set a goal: nationwide adoption of EMR—to include all medical practices—within a decade. In a speech that year at Vanderbilt University Medical Center, the President said: “One of the amazing discrepancies in American society today is we’re literally changing how medicine is delivered in incredibly positive ways, and yet docs are still spending a lot of time writing things on paper.”¹

Certifying body arises from the private sector. Subsequently, the US Department of Health and Human Services (HHS) established the Office of the National Coordinator for Health Information Technology and the American Health Information Community. The sweeping goal of these bodies? Better health care by application of information technology and creation of standards for certifying EMR systems that provide core functionality.

 

In response, 3 private sector health information management groups jointly formed the Certification Commission for Healthcare Information Technology (CCHIT; www.cchit.org). In 2005, this private-sector entity entered into a contract with HHS, to, in the commission’s words, “develop and evaluate certification criteria and create an inspection process for healthcare information technology” in 3 areas:

• Ambulatory EMR for offices
  
• Inpatient EMR for hospitals and health systems
  
• The network components through which EMR share information.
  

The work of this body is ongoing.

Pay-for-performance pushes the issue. Today, insurers—federal and private— are mandating adherence to standards of care for maximal reimbursement of services. These reimbursement schemes, called pay-for-performance, or P4P, are based on providers delivering documentation that specific protocols are followed and outcomes are monitored. The point is that it will be nearly impossible for physicians to comply with insurers’ P4P requirements unless that documentation is in an electronic format.

The market speaks—loudly. Other forces are bringing clinicians to a reckoning with EMR:

• Some malpractice carriers offer a discount on premiums to physicians who document work using EMR
  
• Patients are asking for electronic access to their providers by way of Web sites and e-mail
  
• More and more requests for documentation from multiple interested parties to a patient’s care increase overhead costs and place greater demands on paper-based systems.
  

Physicians cannot meet these demands with paper-based record-keeping.

Reticence has been the watchword

Despite the external and internal forces that are driving adoption, physicians have, as a whole, been reticent to adopt EMR. The nonprofit Healthcare Information and Management Systems Society (HIMSS) reports that 26% of ambulatory practices have adopted EMR, but this penetration is predominantly in multi-specialty clinics and hospital-owned practices.² Few data exist on the penetration of EMR in single-specialty ObGyn practices; anecdotally, vendors estimate a penetration of 10% to 15%.

Why this slow pace toward something broadly acknowledged as key to the well-being of health care?

It means a change. Adopting EMR represents change; well-designed EMR systems streamline workflow in a practice by automating many functions, eliminating duplications of effort, and shifting roles from moving paper to managing digital information. Fear of change and resistance to change are the most common reasons that single-specialty ObGyn practices have not adopted EMR.

It costs. Expense is often cited as the reason why a practice has not adopted an EMR. True: Upfront hardware costs, software costs (license fees, subscription fees), implementation fees, and training costs add up. But a well-designed EMR system should provide a substantial return on investment (ROI) based on savings and on an increase in revenue.

It may be awkward. Some physicians cannot type well. They do not adopt EMR, therefore, because they fear embarrassment using a computer to enter clinical documentation in the consultation room in front of a patient.

On the plus side

On the other side of the coin, the advantages of EMR to physicians are several:

Documentation. EMR facilitate complete documentation of a patient’s visit, current needs and care plan, and record—thereby reducing the clinician’s liability and the risk of medical error. Functions include order entry, prescribing, accurate coding based on work-effort, tracking of outstanding lab tests, and notification.

No chart pulls. With EMR, patient chart pulls are almost nonexistent. A chart is available anywhere a computer is located, any time it is needed.

Decision-making. Probably most importantly, EMR provide clinical decision support by means of alerts (drug interactions, allergies) and reminders (need for follow-up, test orders).

Portal to the patient. Internet-accessed portals that are part of EMR systems facilitate asynchronous communication with patients. A patient can make an appointment, refill a prescription, and request educational materials through such a Web portal. Once an appointment is scheduled, the patient can enter her medical history so that it is specific to the appointment—a feature that is particularly useful when a woman knows the reason that she is visiting the ObGyn (“I’m pregnant,” “My cycle has changed”).

Such a patient-entered history can populate the EMR and contribute elements for appropriate coding. Furthermore, a Web portal in an EMR system enables the physician to reply to a patient with secure messages 1) about lab results, reminders, and appointments and 2) that deliver educational materials.

Keys to embarking on a successful transition

Because EMR are still used by only a small minority of practices, those that seek to move away from the paper record are almost always doing so for the first time. Uncertainty about the adoption adds to anxiety. There are, however, simple steps to take to maintain control over the adoption process and methodically manage it to a successful outcome.

 

Begin with the end in mind. The goal of adoption is not to purchase an EMR system; EMR are only a tool. The goal of the practice should be to transform its existing workflows to make significant improvements over the status quo. Before ever looking at an EMR system, you (and your colleagues, when applicable) must answer several key questions:

• What are we trying to accomplish?
  
• What is it about the status quo that we want to change?
  
• How will we measure success a year after completing the transition?
  

Determine whether you have the resolve to make the transition. As I said, adopting EMR represents change, and the proper motive for adoption is engineered change. Change, however, exposes the human element of transformation. The people who work in the practice are the true determining factor for a successful transforming project, so ask yourself:

• Do you know whether they are ready for change?
  
• Do they understand change?
  
• Are they threatened by it?
  
• Is there broad and vocal leadership backing the impending changes?
  
• Has the impact of the change been discussed with all people involved so they have a clear understanding of its impact on their personal future?
  
• And is the practice, as a team, prepared to go through the turmoil of change as a necessary step on a path to transformation?
  

Assess your sense of urgency—objectively. Because this transition represents a transformation, you’ll have to overcome significant inertia. Without a sense of urgency and aggressive, consistent management of this transition by the leaders of the practice, overcoming human barriers to change will be difficult.

A medical practice that addresses these 3 initial tasks sets itself up for a successful transition from paper to EMR. A good plan—in which goals are well defined and a sense of urgency is consistently communicated and supported by the practice’s leadership—has an excellent chance of resulting in the best possible selection and implementation of an EMR system and accomplishing the goals set for the practice.

In contrast, a transition from paper that begins with such a vague notion as “I guess we need an EMR eventually, so we might as well start now” is much more likely to spark turmoil among staff. The staff then embarks on a selection and implementation process that is heavily influenced by emotion and interpractice politics. They face a diminished opportunity for completing the transition efficiently and successfully.

Throughout this process, the staff should always bear in mind that this is a transformation that they plan, control, and execute. EMR are simply a means to an end—not the end itself.

Two types of systems, various material needs

There are 2 primary configurations of EMR systems: client-server applications and remote-hosted systems. The latter operate through an Application Service Provider (ASP). (See “What are the 2 types of EMR?”)

In addition to type of system, keep these material needs in mind as you plan:

Connectivity. Most medical practices rely on the Internet for a variety of functions; truly, the Internet has become a vital link in health-care information technology.

An ASP system depends on the Internet, whereas client-server applications require a dial-up modem connection or other Internet connectivity to obtain information from outside sources. I recommend purchase of broadband Internet connectivity because it facilitates transmission of large files, such as images and data-rich documents.

Hardware. All EMR require computers for data entry. One attractive option for a medical practice is the TabletPC, which is available from several manufacturers and which uses the Windows XP Tablet PC Edition operating system. Combined with a secure wireless network for moving from room to room, the TabletPC is a technological breakthrough for physicians to document information in a clinical setting. It permits cursive data entry using a special electronic pen, voice recognition entry, and keyboard entry. Whereas a desktop computer places a barrier—the monitor or screen—between physician and patient, a TabletPC emulates the flat, horizontal surface of a paper chart or clipboard.

Importantly, a TabletPC has all the functionality of a desktop computer. Although workflow varies from practice to practice, it can be said generally that most clinical personnel work best with a TabletPC because of its mobility and most clerical personnel work easily with a desktop computer.

The price of a TabletPC? Two to 4 times that of a desktop computer.

Infrastructure. Other devices—printers, scanners, wireless networks, digital cameras—are required to operate an EMR system. A practice that uses a client-server application must purchase a data server. One with a system that operates by remote access either requires a virtual private network (VPN) for secure Internet connection or must install an emulator (such as Citrix).

 

People. Physicians and staff in the practice are always the key to success when implementing an EMR system. Consider a vendor’s ability to assist with the human variables of change management when you assess systems. Even the best EMR cannot save a practice from a poorly planned and executed transformation.

Going shopping

Many physicians want an EMR system to support the conventional process in their practice. This is a prescription for failure! Instead, evaluate the design of an EMR system for its ability to facilitate change in the process and in roles, and to eliminate manual functions and analysis of data.

One EMR system may appear to be the same as the next, but differences are revealed in the way that they optimize workflow. Generic systems may support several primary care medical specialties well, but may impose inefficiencies in other specialties—particularly in niche specialties. Similarly, a specialty-specific system works well for the specialty for which it was designed but is inefficient in another specialty. Approximately 80% of clinical workflow needs may be met by a generic EMR system, but the 20% that are specialty-specific can make the difference between success and failure.

Ensure that the sales presentations you attend address:

• the needs of your practice
  
• the criteria for success that you defined during planning.
  

Require vendors to provide a demonstration that emphasizes the high-frequency daily tasks of your staff, such as documentation of new and return OB patients, annual gynecologic exams combined with gyn problems, collection of histories, and review of lab results. A well-designed EMR, especially one specific to your specialty, should be intuitive and should not disrupt the workflow.

When you arrive at a choice of an EMR system, assess the learning curve that you’ll have to climb for the system to become fully functional. Remember that salesperson? He, or she, gave a slick presentation but you didn’t settle for a dry description; you were sure to try the system live to discover how easily you can learn to navigate its functions.

At last: Implementation and training

EMR vendors develop implementation and training programs to establish their system in a practice and bring it live. These programs are based on vendors’ experience with their product.

As you prepare for your EMR system, keep this in mind: One that’s been well-designed is more than a repository of data. By design, it also reengineers workflow for optimal practice efficiency, safety, and financial management.

Pearls for welcoming a system

Use EMR incrementally. Do not change from paper to electronic abruptly while you, your colleagues, and the staff are learning the system. Vendors of EMR systems design pathways by which physicians ease-in, so to speak, to EMR in a way that minimizes any drop in productivity and loss of revenue. It’s wise to ask vendors about their plan for implementation and training before you sign a contract to purchase.

Expect that it will take 12 to 24 months to convert the practice completely. The time from paper to electronic records depends on the size and age of a practice. In my experience, the half-life of converting an ObGyn practice is approximately 15 months; by 24 months, the records of approximately 95% of active patients will be entered into EMR.

 

Don’t scan all paper records into EMR. Scanning indiscriminately is expensive, disruptive, and doesn’t contribute to ongoing clinical excellence. As part of the conversion process, vendors have methods to enter critical clinical information into the electronic system for uninterrupted use. Instead of wholesale scanning, therefore, be selective and scan only clinically relevant materials—the past several Pap smear results, mammogram reports, operative reports, consultant letters, and similar predefined clinical documents. This usually suffices for ongoing clinical care and avoids excessive expenditure of time, energy, and money.

The promise we talk about needs to become actual

As I noted at the outset, fewer than 25% of physicians have EMR, and estimates are that no more than 10% to 15% of ObGyns have adopted a system. Yet experience has demonstrated: A well-designed EMR offers physicians streamlined workflow, the ability to provide better care, and more time for leisure.

To move the flow and utility of medical information properly into this century, the next step, I urge, is for physicians to recognize the value of EMR, set goals for implementing a system, and reengineer their practice for maximal clinical efficiency, patient safety, and financial gain.

CASE REVISITED: Good outcome; no recurrence

One year later, the ObGyn whose practice was in disarray told a different story: “The ‘Patient Portal’ section of our EMR system is a great time saver. We were amazed at the acceptance and rapid adoption—even our octogenarians love it. The universal access to data is of incalculable value. One of our physicians loves to go home early, have dinner, and then review his charts from home. The EMR improves my recordkeeping, makes encounter documentation more complete, and helps me avoid medication errors. Our billing staff loves the thorough documentation.”

CASE A medical practice in disarray

An ObGyn reported the following signs of a problem to a colleague: “Our practice was literally drowning in paperwork. An exam room was recently converted to hold more charts, and 2 warehouses held our overflow. Employees were constantly searching for records, and telephone messages were delayed for hours or days until the chart could be reviewed. Notoriously bad handwriting and incomplete documentation hampered good communication and good medical care. Transcription costs were out of control. Forms helped but added to the ongoing costs and storage problems.”

What are the treatment options?

Electronic medical records (EMR) have progressed from arcane, slow, cumbersome documentation systems to sophisticated, complex, comprehensive ones. These modern systems hold the potential to reduce administrative and management costs by 30% or more, improve clinical workflow, reduce medical errors, facilitate communication between patient and physician, and enable analysis of data for best practice methods, best outcomes and identifying risks and complications.

For practices like the one described in the preceding paragraph—not a fictional account but actual testimony provided by an ObGyn—EMR offer a powerful potential solution to the problems that result from an overwhelming amount of paper documentation, correspondence, charting, claims, and financial transactions. In this article, I offer a general introduction to EMR; in the next (August) issue of OBG Management, I’ll speak with a group of ObGyns and medical practice managers about their experiences—and inexperience—with EMR.

Progress and paradox

Physicians and scientists have made substantial progress over the past 25 years in pharmacotherapeutics, diagnostic technology, procedures, and treatment protocols. In obstetrics and gynecology alone, consider the array of technologies—3-dimensional ultrasonography, minimally invasive surgery, receptor-specific drugs, in vitro fertilization, long-acting reversible contraceptives—that have advanced the quality and effectiveness of care. Yet little progress has been made in the process of caring for patients.

The fact is that physicians, and other health-care providers, are rooted in paper-based processes that sustain inefficiencies, increase costs, and defy the gains that other industries have made by adopting electronic technologies for handling information. Why are we so stuck?

The state of EMR

EMR—of varying functionality—have been available for longer than 20 years. Early models were developed by physicians who had an interest in software coding and design, and were of limited functionality, arcane, and difficult to use in a clinical setting. Some of those early models, and even a few commercial systems in use today, rely on scanning paper documents into computer files. Such systems may eliminate some paper and facilitate document retrieval, but they do nothing to ease management of the complex transactions of health care, and they do not address handwriting illegibility.

Development of complex EMR systems was limited by primitive technology, inadequate distribution channels, and programming that was cumbersome and expensive to maintain. But these barriers have been overcome with fast processors, inexpensive and abundant memory, broadband Internet connectivity, and programming languages that facilitate automated software development.

Modern EMR systems are not simply data repositories: They also support workflow from the beginning to the end of a patient’s consultation with a health-care provider—an event that generates multiple transactions with multiple recipients. A single consultation may, for example, generate orders for lab tests, imaging studies, a surgical procedure, consultation with other physicians, prescriptions, and counseling, and record the subsequent financial transaction. EMR systems by necessity interact with multiple organizations, institutions, instruments, and other software systems. To software developers, and to the clinicians who use their systems, the challenge is to deftly navigate the complexities of health care.

Forces accelerating adoption

Momentum from the Executive Office. In 2004, President George W. Bush set a goal: nationwide adoption of EMR—to include all medical practices—within a decade. In a speech that year at Vanderbilt University Medical Center, the President said: “One of the amazing discrepancies in American society today is we’re literally changing how medicine is delivered in incredibly positive ways, and yet docs are still spending a lot of time writing things on paper.”¹

Certifying body arises from the private sector. Subsequently, the US Department of Health and Human Services (HHS) established the Office of the National Coordinator for Health Information Technology and the American Health Information Community. The sweeping goal of these bodies? Better health care by application of information technology and creation of standards for certifying EMR systems that provide core functionality.

 

In response, 3 private sector health information management groups jointly formed the Certification Commission for Healthcare Information Technology (CCHIT; www.cchit.org). In 2005, this private-sector entity entered into a contract with HHS, to, in the commission’s words, “develop and evaluate certification criteria and create an inspection process for healthcare information technology” in 3 areas:

• Ambulatory EMR for offices
  
• Inpatient EMR for hospitals and health systems
  
• The network components through which EMR share information.
  

The work of this body is ongoing.

Pay-for-performance pushes the issue. Today, insurers—federal and private— are mandating adherence to standards of care for maximal reimbursement of services. These reimbursement schemes, called pay-for-performance, or P4P, are based on providers delivering documentation that specific protocols are followed and outcomes are monitored. The point is that it will be nearly impossible for physicians to comply with insurers’ P4P requirements unless that documentation is in an electronic format.

The market speaks—loudly. Other forces are bringing clinicians to a reckoning with EMR:

• Some malpractice carriers offer a discount on premiums to physicians who document work using EMR
  
• Patients are asking for electronic access to their providers by way of Web sites and e-mail
  
• More and more requests for documentation from multiple interested parties to a patient’s care increase overhead costs and place greater demands on paper-based systems.
  

Physicians cannot meet these demands with paper-based record-keeping.

Reticence has been the watchword

Despite the external and internal forces that are driving adoption, physicians have, as a whole, been reticent to adopt EMR. The nonprofit Healthcare Information and Management Systems Society (HIMSS) reports that 26% of ambulatory practices have adopted EMR, but this penetration is predominantly in multi-specialty clinics and hospital-owned practices.² Few data exist on the penetration of EMR in single-specialty ObGyn practices; anecdotally, vendors estimate a penetration of 10% to 15%.

Why this slow pace toward something broadly acknowledged as key to the well-being of health care?

It means a change. Adopting EMR represents change; well-designed EMR systems streamline workflow in a practice by automating many functions, eliminating duplications of effort, and shifting roles from moving paper to managing digital information. Fear of change and resistance to change are the most common reasons that single-specialty ObGyn practices have not adopted EMR.

It costs. Expense is often cited as the reason why a practice has not adopted an EMR. True: Upfront hardware costs, software costs (license fees, subscription fees), implementation fees, and training costs add up. But a well-designed EMR system should provide a substantial return on investment (ROI) based on savings and on an increase in revenue.

It may be awkward. Some physicians cannot type well. They do not adopt EMR, therefore, because they fear embarrassment using a computer to enter clinical documentation in the consultation room in front of a patient.

On the plus side

On the other side of the coin, the advantages of EMR to physicians are several:

Documentation. EMR facilitate complete documentation of a patient’s visit, current needs and care plan, and record—thereby reducing the clinician’s liability and the risk of medical error. Functions include order entry, prescribing, accurate coding based on work-effort, tracking of outstanding lab tests, and notification.

No chart pulls. With EMR, patient chart pulls are almost nonexistent. A chart is available anywhere a computer is located, any time it is needed.

Decision-making. Probably most importantly, EMR provide clinical decision support by means of alerts (drug interactions, allergies) and reminders (need for follow-up, test orders).

Portal to the patient. Internet-accessed portals that are part of EMR systems facilitate asynchronous communication with patients. A patient can make an appointment, refill a prescription, and request educational materials through such a Web portal. Once an appointment is scheduled, the patient can enter her medical history so that it is specific to the appointment—a feature that is particularly useful when a woman knows the reason that she is visiting the ObGyn (“I’m pregnant,” “My cycle has changed”).

Such a patient-entered history can populate the EMR and contribute elements for appropriate coding. Furthermore, a Web portal in an EMR system enables the physician to reply to a patient with secure messages 1) about lab results, reminders, and appointments and 2) that deliver educational materials.

Keys to embarking on a successful transition

Because EMR are still used by only a small minority of practices, those that seek to move away from the paper record are almost always doing so for the first time. Uncertainty about the adoption adds to anxiety. There are, however, simple steps to take to maintain control over the adoption process and methodically manage it to a successful outcome.

 

Begin with the end in mind. The goal of adoption is not to purchase an EMR system; EMR are only a tool. The goal of the practice should be to transform its existing workflows to make significant improvements over the status quo. Before ever looking at an EMR system, you (and your colleagues, when applicable) must answer several key questions:

• What are we trying to accomplish?
  
• What is it about the status quo that we want to change?
  
• How will we measure success a year after completing the transition?
  

Determine whether you have the resolve to make the transition. As I said, adopting EMR represents change, and the proper motive for adoption is engineered change. Change, however, exposes the human element of transformation. The people who work in the practice are the true determining factor for a successful transforming project, so ask yourself:

• Do you know whether they are ready for change?
  
• Do they understand change?
  
• Are they threatened by it?
  
• Is there broad and vocal leadership backing the impending changes?
  
• Has the impact of the change been discussed with all people involved so they have a clear understanding of its impact on their personal future?
  
• And is the practice, as a team, prepared to go through the turmoil of change as a necessary step on a path to transformation?
  

Assess your sense of urgency—objectively. Because this transition represents a transformation, you’ll have to overcome significant inertia. Without a sense of urgency and aggressive, consistent management of this transition by the leaders of the practice, overcoming human barriers to change will be difficult.

A medical practice that addresses these 3 initial tasks sets itself up for a successful transition from paper to EMR. A good plan—in which goals are well defined and a sense of urgency is consistently communicated and supported by the practice’s leadership—has an excellent chance of resulting in the best possible selection and implementation of an EMR system and accomplishing the goals set for the practice.

In contrast, a transition from paper that begins with such a vague notion as “I guess we need an EMR eventually, so we might as well start now” is much more likely to spark turmoil among staff. The staff then embarks on a selection and implementation process that is heavily influenced by emotion and interpractice politics. They face a diminished opportunity for completing the transition efficiently and successfully.

Throughout this process, the staff should always bear in mind that this is a transformation that they plan, control, and execute. EMR are simply a means to an end—not the end itself.

Two types of systems, various material needs

There are 2 primary configurations of EMR systems: client-server applications and remote-hosted systems. The latter operate through an Application Service Provider (ASP). (See “What are the 2 types of EMR?”)

In addition to type of system, keep these material needs in mind as you plan:

Connectivity. Most medical practices rely on the Internet for a variety of functions; truly, the Internet has become a vital link in health-care information technology.

An ASP system depends on the Internet, whereas client-server applications require a dial-up modem connection or other Internet connectivity to obtain information from outside sources. I recommend purchase of broadband Internet connectivity because it facilitates transmission of large files, such as images and data-rich documents.

Hardware. All EMR require computers for data entry. One attractive option for a medical practice is the TabletPC, which is available from several manufacturers and which uses the Windows XP Tablet PC Edition operating system. Combined with a secure wireless network for moving from room to room, the TabletPC is a technological breakthrough for physicians to document information in a clinical setting. It permits cursive data entry using a special electronic pen, voice recognition entry, and keyboard entry. Whereas a desktop computer places a barrier—the monitor or screen—between physician and patient, a TabletPC emulates the flat, horizontal surface of a paper chart or clipboard.

Importantly, a TabletPC has all the functionality of a desktop computer. Although workflow varies from practice to practice, it can be said generally that most clinical personnel work best with a TabletPC because of its mobility and most clerical personnel work easily with a desktop computer.

The price of a TabletPC? Two to 4 times that of a desktop computer.

Infrastructure. Other devices—printers, scanners, wireless networks, digital cameras—are required to operate an EMR system. A practice that uses a client-server application must purchase a data server. One with a system that operates by remote access either requires a virtual private network (VPN) for secure Internet connection or must install an emulator (such as Citrix).

 

People. Physicians and staff in the practice are always the key to success when implementing an EMR system. Consider a vendor’s ability to assist with the human variables of change management when you assess systems. Even the best EMR cannot save a practice from a poorly planned and executed transformation.

Going shopping

Many physicians want an EMR system to support the conventional process in their practice. This is a prescription for failure! Instead, evaluate the design of an EMR system for its ability to facilitate change in the process and in roles, and to eliminate manual functions and analysis of data.

One EMR system may appear to be the same as the next, but differences are revealed in the way that they optimize workflow. Generic systems may support several primary care medical specialties well, but may impose inefficiencies in other specialties—particularly in niche specialties. Similarly, a specialty-specific system works well for the specialty for which it was designed but is inefficient in another specialty. Approximately 80% of clinical workflow needs may be met by a generic EMR system, but the 20% that are specialty-specific can make the difference between success and failure.

Ensure that the sales presentations you attend address:

• the needs of your practice
  
• the criteria for success that you defined during planning.
  

Require vendors to provide a demonstration that emphasizes the high-frequency daily tasks of your staff, such as documentation of new and return OB patients, annual gynecologic exams combined with gyn problems, collection of histories, and review of lab results. A well-designed EMR, especially one specific to your specialty, should be intuitive and should not disrupt the workflow.

When you arrive at a choice of an EMR system, assess the learning curve that you’ll have to climb for the system to become fully functional. Remember that salesperson? He, or she, gave a slick presentation but you didn’t settle for a dry description; you were sure to try the system live to discover how easily you can learn to navigate its functions.

At last: Implementation and training

EMR vendors develop implementation and training programs to establish their system in a practice and bring it live. These programs are based on vendors’ experience with their product.

As you prepare for your EMR system, keep this in mind: One that’s been well-designed is more than a repository of data. By design, it also reengineers workflow for optimal practice efficiency, safety, and financial management.

Pearls for welcoming a system

Use EMR incrementally. Do not change from paper to electronic abruptly while you, your colleagues, and the staff are learning the system. Vendors of EMR systems design pathways by which physicians ease-in, so to speak, to EMR in a way that minimizes any drop in productivity and loss of revenue. It’s wise to ask vendors about their plan for implementation and training before you sign a contract to purchase.

Expect that it will take 12 to 24 months to convert the practice completely. The time from paper to electronic records depends on the size and age of a practice. In my experience, the half-life of converting an ObGyn practice is approximately 15 months; by 24 months, the records of approximately 95% of active patients will be entered into EMR.

 

Don’t scan all paper records into EMR. Scanning indiscriminately is expensive, disruptive, and doesn’t contribute to ongoing clinical excellence. As part of the conversion process, vendors have methods to enter critical clinical information into the electronic system for uninterrupted use. Instead of wholesale scanning, therefore, be selective and scan only clinically relevant materials—the past several Pap smear results, mammogram reports, operative reports, consultant letters, and similar predefined clinical documents. This usually suffices for ongoing clinical care and avoids excessive expenditure of time, energy, and money.

The promise we talk about needs to become actual

As I noted at the outset, fewer than 25% of physicians have EMR, and estimates are that no more than 10% to 15% of ObGyns have adopted a system. Yet experience has demonstrated: A well-designed EMR offers physicians streamlined workflow, the ability to provide better care, and more time for leisure.

To move the flow and utility of medical information properly into this century, the next step, I urge, is for physicians to recognize the value of EMR, set goals for implementing a system, and reengineer their practice for maximal clinical efficiency, patient safety, and financial gain.

CASE REVISITED: Good outcome; no recurrence

One year later, the ObGyn whose practice was in disarray told a different story: “The ‘Patient Portal’ section of our EMR system is a great time saver. We were amazed at the acceptance and rapid adoption—even our octogenarians love it. The universal access to data is of incalculable value. One of our physicians loves to go home early, have dinner, and then review his charts from home. The EMR improves my recordkeeping, makes encounter documentation more complete, and helps me avoid medication errors. Our billing staff loves the thorough documentation.”

CASE A medical practice in disarray

An ObGyn reported the following signs of a problem to a colleague: “Our practice was literally drowning in paperwork. An exam room was recently converted to hold more charts, and 2 warehouses held our overflow. Employees were constantly searching for records, and telephone messages were delayed for hours or days until the chart could be reviewed. Notoriously bad handwriting and incomplete documentation hampered good communication and good medical care. Transcription costs were out of control. Forms helped but added to the ongoing costs and storage problems.”

What are the treatment options?

Electronic medical records (EMR) have progressed from arcane, slow, cumbersome documentation systems to sophisticated, complex, comprehensive ones. These modern systems hold the potential to reduce administrative and management costs by 30% or more, improve clinical workflow, reduce medical errors, facilitate communication between patient and physician, and enable analysis of data for best practice methods, best outcomes and identifying risks and complications.

For practices like the one described in the preceding paragraph—not a fictional account but actual testimony provided by an ObGyn—EMR offer a powerful potential solution to the problems that result from an overwhelming amount of paper documentation, correspondence, charting, claims, and financial transactions. In this article, I offer a general introduction to EMR; in the next (August) issue of OBG Management, I’ll speak with a group of ObGyns and medical practice managers about their experiences—and inexperience—with EMR.

Progress and paradox

Physicians and scientists have made substantial progress over the past 25 years in pharmacotherapeutics, diagnostic technology, procedures, and treatment protocols. In obstetrics and gynecology alone, consider the array of technologies—3-dimensional ultrasonography, minimally invasive surgery, receptor-specific drugs, in vitro fertilization, long-acting reversible contraceptives—that have advanced the quality and effectiveness of care. Yet little progress has been made in the process of caring for patients.

The fact is that physicians, and other health-care providers, are rooted in paper-based processes that sustain inefficiencies, increase costs, and defy the gains that other industries have made by adopting electronic technologies for handling information. Why are we so stuck?

The state of EMR

EMR—of varying functionality—have been available for longer than 20 years. Early models were developed by physicians who had an interest in software coding and design, and were of limited functionality, arcane, and difficult to use in a clinical setting. Some of those early models, and even a few commercial systems in use today, rely on scanning paper documents into computer files. Such systems may eliminate some paper and facilitate document retrieval, but they do nothing to ease management of the complex transactions of health care, and they do not address handwriting illegibility.

Development of complex EMR systems was limited by primitive technology, inadequate distribution channels, and programming that was cumbersome and expensive to maintain. But these barriers have been overcome with fast processors, inexpensive and abundant memory, broadband Internet connectivity, and programming languages that facilitate automated software development.

Modern EMR systems are not simply data repositories: They also support workflow from the beginning to the end of a patient’s consultation with a health-care provider—an event that generates multiple transactions with multiple recipients. A single consultation may, for example, generate orders for lab tests, imaging studies, a surgical procedure, consultation with other physicians, prescriptions, and counseling, and record the subsequent financial transaction. EMR systems by necessity interact with multiple organizations, institutions, instruments, and other software systems. To software developers, and to the clinicians who use their systems, the challenge is to deftly navigate the complexities of health care.

Forces accelerating adoption

Momentum from the Executive Office. In 2004, President George W. Bush set a goal: nationwide adoption of EMR—to include all medical practices—within a decade. In a speech that year at Vanderbilt University Medical Center, the President said: “One of the amazing discrepancies in American society today is we’re literally changing how medicine is delivered in incredibly positive ways, and yet docs are still spending a lot of time writing things on paper.”¹

Certifying body arises from the private sector. Subsequently, the US Department of Health and Human Services (HHS) established the Office of the National Coordinator for Health Information Technology and the American Health Information Community. The sweeping goal of these bodies? Better health care by application of information technology and creation of standards for certifying EMR systems that provide core functionality.

 

In response, 3 private sector health information management groups jointly formed the Certification Commission for Healthcare Information Technology (CCHIT; www.cchit.org). In 2005, this private-sector entity entered into a contract with HHS, to, in the commission’s words, “develop and evaluate certification criteria and create an inspection process for healthcare information technology” in 3 areas:

• Ambulatory EMR for offices
  
• Inpatient EMR for hospitals and health systems
  
• The network components through which EMR share information.
  

The work of this body is ongoing.

Pay-for-performance pushes the issue. Today, insurers—federal and private— are mandating adherence to standards of care for maximal reimbursement of services. These reimbursement schemes, called pay-for-performance, or P4P, are based on providers delivering documentation that specific protocols are followed and outcomes are monitored. The point is that it will be nearly impossible for physicians to comply with insurers’ P4P requirements unless that documentation is in an electronic format.

The market speaks—loudly. Other forces are bringing clinicians to a reckoning with EMR:

• Some malpractice carriers offer a discount on premiums to physicians who document work using EMR
  
• Patients are asking for electronic access to their providers by way of Web sites and e-mail
  
• More and more requests for documentation from multiple interested parties to a patient’s care increase overhead costs and place greater demands on paper-based systems.
  

Physicians cannot meet these demands with paper-based record-keeping.

Reticence has been the watchword

Despite the external and internal forces that are driving adoption, physicians have, as a whole, been reticent to adopt EMR. The nonprofit Healthcare Information and Management Systems Society (HIMSS) reports that 26% of ambulatory practices have adopted EMR, but this penetration is predominantly in multi-specialty clinics and hospital-owned practices.² Few data exist on the penetration of EMR in single-specialty ObGyn practices; anecdotally, vendors estimate a penetration of 10% to 15%.

Why this slow pace toward something broadly acknowledged as key to the well-being of health care?

It means a change. Adopting EMR represents change; well-designed EMR systems streamline workflow in a practice by automating many functions, eliminating duplications of effort, and shifting roles from moving paper to managing digital information. Fear of change and resistance to change are the most common reasons that single-specialty ObGyn practices have not adopted EMR.

It costs. Expense is often cited as the reason why a practice has not adopted an EMR. True: Upfront hardware costs, software costs (license fees, subscription fees), implementation fees, and training costs add up. But a well-designed EMR system should provide a substantial return on investment (ROI) based on savings and on an increase in revenue.

It may be awkward. Some physicians cannot type well. They do not adopt EMR, therefore, because they fear embarrassment using a computer to enter clinical documentation in the consultation room in front of a patient.

On the plus side

On the other side of the coin, the advantages of EMR to physicians are several:

Documentation. EMR facilitate complete documentation of a patient’s visit, current needs and care plan, and record—thereby reducing the clinician’s liability and the risk of medical error. Functions include order entry, prescribing, accurate coding based on work-effort, tracking of outstanding lab tests, and notification.

No chart pulls. With EMR, patient chart pulls are almost nonexistent. A chart is available anywhere a computer is located, any time it is needed.

Decision-making. Probably most importantly, EMR provide clinical decision support by means of alerts (drug interactions, allergies) and reminders (need for follow-up, test orders).

Portal to the patient. Internet-accessed portals that are part of EMR systems facilitate asynchronous communication with patients. A patient can make an appointment, refill a prescription, and request educational materials through such a Web portal. Once an appointment is scheduled, the patient can enter her medical history so that it is specific to the appointment—a feature that is particularly useful when a woman knows the reason that she is visiting the ObGyn (“I’m pregnant,” “My cycle has changed”).

Such a patient-entered history can populate the EMR and contribute elements for appropriate coding. Furthermore, a Web portal in an EMR system enables the physician to reply to a patient with secure messages 1) about lab results, reminders, and appointments and 2) that deliver educational materials.

Keys to embarking on a successful transition

Because EMR are still used by only a small minority of practices, those that seek to move away from the paper record are almost always doing so for the first time. Uncertainty about the adoption adds to anxiety. There are, however, simple steps to take to maintain control over the adoption process and methodically manage it to a successful outcome.

 

Begin with the end in mind. The goal of adoption is not to purchase an EMR system; EMR are only a tool. The goal of the practice should be to transform its existing workflows to make significant improvements over the status quo. Before ever looking at an EMR system, you (and your colleagues, when applicable) must answer several key questions:

• What are we trying to accomplish?
  
• What is it about the status quo that we want to change?
  
• How will we measure success a year after completing the transition?
  

Determine whether you have the resolve to make the transition. As I said, adopting EMR represents change, and the proper motive for adoption is engineered change. Change, however, exposes the human element of transformation. The people who work in the practice are the true determining factor for a successful transforming project, so ask yourself:

• Do you know whether they are ready for change?
  
• Do they understand change?
  
• Are they threatened by it?
  
• Is there broad and vocal leadership backing the impending changes?
  
• Has the impact of the change been discussed with all people involved so they have a clear understanding of its impact on their personal future?
  
• And is the practice, as a team, prepared to go through the turmoil of change as a necessary step on a path to transformation?
  

Assess your sense of urgency—objectively. Because this transition represents a transformation, you’ll have to overcome significant inertia. Without a sense of urgency and aggressive, consistent management of this transition by the leaders of the practice, overcoming human barriers to change will be difficult.

A medical practice that addresses these 3 initial tasks sets itself up for a successful transition from paper to EMR. A good plan—in which goals are well defined and a sense of urgency is consistently communicated and supported by the practice’s leadership—has an excellent chance of resulting in the best possible selection and implementation of an EMR system and accomplishing the goals set for the practice.

In contrast, a transition from paper that begins with such a vague notion as “I guess we need an EMR eventually, so we might as well start now” is much more likely to spark turmoil among staff. The staff then embarks on a selection and implementation process that is heavily influenced by emotion and interpractice politics. They face a diminished opportunity for completing the transition efficiently and successfully.

Throughout this process, the staff should always bear in mind that this is a transformation that they plan, control, and execute. EMR are simply a means to an end—not the end itself.

Two types of systems, various material needs

There are 2 primary configurations of EMR systems: client-server applications and remote-hosted systems. The latter operate through an Application Service Provider (ASP). (See “What are the 2 types of EMR?”)

In addition to type of system, keep these material needs in mind as you plan:

Connectivity. Most medical practices rely on the Internet for a variety of functions; truly, the Internet has become a vital link in health-care information technology.

An ASP system depends on the Internet, whereas client-server applications require a dial-up modem connection or other Internet connectivity to obtain information from outside sources. I recommend purchase of broadband Internet connectivity because it facilitates transmission of large files, such as images and data-rich documents.

Hardware. All EMR require computers for data entry. One attractive option for a medical practice is the TabletPC, which is available from several manufacturers and which uses the Windows XP Tablet PC Edition operating system. Combined with a secure wireless network for moving from room to room, the TabletPC is a technological breakthrough for physicians to document information in a clinical setting. It permits cursive data entry using a special electronic pen, voice recognition entry, and keyboard entry. Whereas a desktop computer places a barrier—the monitor or screen—between physician and patient, a TabletPC emulates the flat, horizontal surface of a paper chart or clipboard.

Importantly, a TabletPC has all the functionality of a desktop computer. Although workflow varies from practice to practice, it can be said generally that most clinical personnel work best with a TabletPC because of its mobility and most clerical personnel work easily with a desktop computer.

The price of a TabletPC? Two to 4 times that of a desktop computer.

Infrastructure. Other devices—printers, scanners, wireless networks, digital cameras—are required to operate an EMR system. A practice that uses a client-server application must purchase a data server. One with a system that operates by remote access either requires a virtual private network (VPN) for secure Internet connection or must install an emulator (such as Citrix).

 

People. Physicians and staff in the practice are always the key to success when implementing an EMR system. Consider a vendor’s ability to assist with the human variables of change management when you assess systems. Even the best EMR cannot save a practice from a poorly planned and executed transformation.

Going shopping

Many physicians want an EMR system to support the conventional process in their practice. This is a prescription for failure! Instead, evaluate the design of an EMR system for its ability to facilitate change in the process and in roles, and to eliminate manual functions and analysis of data.

One EMR system may appear to be the same as the next, but differences are revealed in the way that they optimize workflow. Generic systems may support several primary care medical specialties well, but may impose inefficiencies in other specialties—particularly in niche specialties. Similarly, a specialty-specific system works well for the specialty for which it was designed but is inefficient in another specialty. Approximately 80% of clinical workflow needs may be met by a generic EMR system, but the 20% that are specialty-specific can make the difference between success and failure.

Ensure that the sales presentations you attend address:

• the needs of your practice
  
• the criteria for success that you defined during planning.
  

Require vendors to provide a demonstration that emphasizes the high-frequency daily tasks of your staff, such as documentation of new and return OB patients, annual gynecologic exams combined with gyn problems, collection of histories, and review of lab results. A well-designed EMR, especially one specific to your specialty, should be intuitive and should not disrupt the workflow.

When you arrive at a choice of an EMR system, assess the learning curve that you’ll have to climb for the system to become fully functional. Remember that salesperson? He, or she, gave a slick presentation but you didn’t settle for a dry description; you were sure to try the system live to discover how easily you can learn to navigate its functions.

At last: Implementation and training

EMR vendors develop implementation and training programs to establish their system in a practice and bring it live. These programs are based on vendors’ experience with their product.

As you prepare for your EMR system, keep this in mind: One that’s been well-designed is more than a repository of data. By design, it also reengineers workflow for optimal practice efficiency, safety, and financial management.

Pearls for welcoming a system

Use EMR incrementally. Do not change from paper to electronic abruptly while you, your colleagues, and the staff are learning the system. Vendors of EMR systems design pathways by which physicians ease-in, so to speak, to EMR in a way that minimizes any drop in productivity and loss of revenue. It’s wise to ask vendors about their plan for implementation and training before you sign a contract to purchase.

Expect that it will take 12 to 24 months to convert the practice completely. The time from paper to electronic records depends on the size and age of a practice. In my experience, the half-life of converting an ObGyn practice is approximately 15 months; by 24 months, the records of approximately 95% of active patients will be entered into EMR.

 

Don’t scan all paper records into EMR. Scanning indiscriminately is expensive, disruptive, and doesn’t contribute to ongoing clinical excellence. As part of the conversion process, vendors have methods to enter critical clinical information into the electronic system for uninterrupted use. Instead of wholesale scanning, therefore, be selective and scan only clinically relevant materials—the past several Pap smear results, mammogram reports, operative reports, consultant letters, and similar predefined clinical documents. This usually suffices for ongoing clinical care and avoids excessive expenditure of time, energy, and money.

The promise we talk about needs to become actual

As I noted at the outset, fewer than 25% of physicians have EMR, and estimates are that no more than 10% to 15% of ObGyns have adopted a system. Yet experience has demonstrated: A well-designed EMR offers physicians streamlined workflow, the ability to provide better care, and more time for leisure.

To move the flow and utility of medical information properly into this century, the next step, I urge, is for physicians to recognize the value of EMR, set goals for implementing a system, and reengineer their practice for maximal clinical efficiency, patient safety, and financial gain.

CASE REVISITED: Good outcome; no recurrence

One year later, the ObGyn whose practice was in disarray told a different story: “The ‘Patient Portal’ section of our EMR system is a great time saver. We were amazed at the acceptance and rapid adoption—even our octogenarians love it. The universal access to data is of incalculable value. One of our physicians loves to go home early, have dinner, and then review his charts from home. The EMR improves my recordkeeping, makes encounter documentation more complete, and helps me avoid medication errors. Our billing staff loves the thorough documentation.”

ObGyns perform most of the screening for cancers of the ovary and breast. The first cancer is especially lethal, though rare, and the second is especially feared among women. This update reviews screening guidelines and recent studies that may affect how we detect and prevent ovarian and breast cancers.

Among the findings:

• In the only multicenter, prospective, randomized, controlled study to date to look at the use of CA-125 and transvaginal ultrasound screening in a low-risk population of postmenopausal women in the United States, researchers found no evidence to suggest a need to revise the present (1996) ovarian cancer screening guidelines of the US Preventive Services Task Force.
• Using a Markov decision-analysis model, investigators explored the health effects of prophylactic bilateral salpingo-oophorectomy in women at average risk of ovarian cancer undergoing hysterectomy. They found that removing the ovaries may decrease overall survival.
• Investigators found the opposite to be true in women with BRCA1 or BRCA2 mutations. Prophylactic bilateral salpingo-oophorectomy greatly reduced the overall mortality rate, as well as the risk of ovarian and breast cancer.
• In a prospective cohort study of BRCA mutation carriers with no history of breast cancer who underwent prophylactic oophorectomy, researchers found the short-term use of hormone replacement therapy to be safe, with no loss of protection against breast cancer.

No need to revise screening guidelines for ovarian cancer

Buys S, Partridge E, Greene M, et al; for the PLCO Project Team. Ovarian cancer screening in the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial: findings from the initial screen of a randomized trial. Am J Obstet Gynecol. 2005;193:1630–1639.

The need to identify a marker for the early detection of ovarian cancer is especially urgent, given that approximately 75% of women with the cancer present with late-stage disease. Because the disease is rare, finding a cost-effective screening test with good sensitivity and very high specificity (to decrease too many false-positive results) will be challenging.

So far, no prospective, randomized studies of any ovarian cancer screening modality have demonstrated a decrease in mortality—the gold standard of efficacy for any screening test. Therefore, the Prostate, Lung, Colorectal, and Ovarian (PLCO) cancer trial is a critical study—it is the only multicenter, prospective, randomized, controlled study in the United States to tackle the question of whether CA-125 and transvaginal ultrasonographic (US) screening will be effective in a low-risk population of postmenopausal women aged 55 to 74.

In this large study, 1 arm underwent ovarian cancer screening with both modalities and the other arm underwent no such screening.

This study reports on baseline, or ‘prevalent,’ cancers

This preliminary report does not comment on the efficacy of ovarian cancer screening; data on the effect of repeated annual screens on detection rates and mortality will become available over the next several years.

Rather, the purpose of this preliminary report was to detail the baseline ovarian cancer screening tests of the 39,115 women randomized to the intervention arm from November 15, 1993, to December 13, 2001. These results describe “prevalent” cancers—that is, cancers that are present on the first screen. The more important information about efficacy of screening will come over the next several years, as “incident” cancers develop.

Roughly 6% of women had at least 1 abnormal finding at baseline

Among 28,506 women with results for both baseline tests, 1,706 had at least 1 abnormal finding:

• 1,338 had an abnormal transvaginal US scan
• 402 had an abnormal level on the CA-125 test
• 34 had abnormalities in both tests
• 29 malignant neoplasms were identified in this population, 20 of them invasive.

When combined, CA-125 and transvaginal ultrasonography had good positive predictive value

In general, a positive predictive value (PPV) of more than 10% (ie, 10 surgeries to detect 1 cancer) is considered reasonable justification for a screening test. In the PLCO trial, the PPV was 4% for CA-125 alone (16 neoplasms in 402 positive screens), 1.6% for transvaginal ultrasonography alone (22 neoplasms in 1,338 positive screens), and 26.5% if both tests were abnormal (9 neoplasms in 34 positive screens).

When tumors of low malignant potential were excluded, the PPV was 3.7% for an abnormal CA-125, 1.0% for an abnormal transvaginal sonogram, and 23.5% if both tests were abnormal. A PPV of 23.5% for both tests is fairly good (ie, approximately 4 surgeries to detect 1 cancer). However, if only women in whom both screening tests were abnormal went to surgery, 12 of 20 invasive cancers would be missed.

 

Bottom line: Routine screening still not justified

Nothing in the findings reported here suggests that we need to revise the current (from 1996) ovarian cancer screening guidelines of the US Preventive Services Task Force,¹ which state that “routine screening for ovarian cancer by US, the measurement of serum tumor markers, or pelvic examination is not recommended.”

We will need to wait until the PLCO trial results come in to see the effect of repeated annual ovarian cancer screens on detection rates and mortality.

Consider ovarian conservation in hysterectomy for benign disease

Parker W, Broder MS, Liu Z, Shoupe D, Farquhar C, Berek JS. Ovarian conservation at the time of hysterectomy for benign disease. Obstet Gynecol. 2005;106:219–226.

As discussed earlier, we have no good screening tool for the early detection of ovarian cancer. Although rare, ovarian cancer is a lethal, scary disease, and most ObGyns prophylactically remove the ovaries at the time of hysterectomy in most postmenopausal and many perimenopausal women.

The downside to this strategy seems low among postmenopausal women, and the upside, in terms of not having to worry about ovarian cancer, seems high. The study by Parker and colleagues, while having definite limitations, asks us to question this routine practice pattern. The authors found that prophylactic oophorectomy may be associated with decreased overall survival.

Model used SEER data, Nurses’ Health Study to predict survival

Parker and colleagues used a Markov decision-analysis model (a hypothetical mathematical model that uses published data to create cohorts of patients to estimate risk of morbidity or mortality, or both, over time) to evaluate the risks and benefits of ovarian conservation at the time of hysterectomy for benign disease. Age-specific mortality estimates for ovarian cancer were based on Surveillance, Epidemiology and End Results (SEER) statistics.

For women at average risk of ovarian cancer, the probability of surviving to 80 years of age after hysterectomy between 50 and 54 years varied, and was 62.8% and 62.5% for ovarian conservation with and without estrogen therapy, respectively, compared with 62.2% and 53.9% for oophorectomy with and without estrogen therapy. The main reason that the model found decreased overall survival with prophylactic oophorectomy was an increase in coronary artery disease after oophorectomy—a finding that was based on data from the Nurses’ Health Study.

At the very least, think hard about the decision to remove the ovaries

This report estimated that about 300,000 prophylactic oophorectomies are carried out annually in the United States. Although this study has limitations, we believe it encourages debate and reexamination of the benefit of prophylactic oophorectomy for benign indications in young, low-risk patients.

The most important finding from the study is that oophorectomy conferred no survival advantage. Given the rarity of ovarian cancer among the general population, this effect is not that surprising.

For now, careful risk assessment remains a fundamental component of management, so that women who are at increased risk of ovarian cancer can undergo prophylactic salpingo-oophorectomy.

For low-risk women, who constitute the majority of patients, the data to support removing ovaries at the time of hysterectomy are less clear.

Make sure the patient understands the low risk of cancer and possible cardiac benefits of preservation

• Conduct a thorough discussion with the patient about the pros and cons of oophorectomy for benign disease.
• The risk of ovarian cancer in the general population is low; patients should understand that they may derive cardiac protection from postmenopausal ovarian function.

Do consider oophorectomy among carriers of a BRCA mutation

Domchek SM, Friebel TM, Neuhausen SL, et al. Mortality after bilateral salpingo-oophorectomy in BRCA1 and BRCA2 mutation carriers: a prospective cohort study. Lancet Oncol. 2006;7:223–229.

Women known to have BRCA1 or BRCA2 mutations are now managed by means of surveillance or with prophylactic bilateral salpingo-oophorectomy. Although bilateral salpingo-oophorectomy has been shown to reduce the risk of ovarian cancer by 90% and the risk of breast cancer by 50%, until this study few data shed light on the effect of the procedure on overall mortality among women with BRCA mutations.

This prospective cohort study identified 155 patients with BRCA1 or BRCA2 mutations who elected to undergo bilateral salpingo-oophorectomy and matched them by age with a control group of 271. All women were followed until death by any cause or by breast, ovarian, or primary peritoneal cancer. The women were followed for a mean of 3.1 years in the bilateral salpingo-oophorectomy group and 2.1 years in the control group.

Overall and cancer-specific survival improved with oophorectomy

Among BRCA mutation carriers, women who chose prophylactic bilateral salpingo-oophorectomy had improved overall and cancer-specific survival, compared with women who did not undergo the surgery.

 

In the analysis of the matched BRCA mutation carriers, women who chose to undergo the procedure had a decreased risk of overall mortality (hazard ratio [HR]=0.24; 95% confidence interval [CI], 0.08–0.71); they also had a decreased risk of mortality due to both breast cancer (HR=0.1; 95% CI, 0.02–0.71) and ovarian cancer (HR=0.05; 95% CI, 0.01–0.46).

Practice recommendations

Apparently, unlike women at average risk of ovarian cancer (for whom prophylactic oophorectomy in conjunction with hysterectomy for benign disease may be associated with decreased overall survival; see the review of the study by Parker and colleagues), women with BRCA mutations may benefit from oophorectomy.

Advantages of prophylactic bilateral salpingo-oophorectomy in this patient population should be discussed with potential surgical candidates, because:

• Women who have a BRCA mutation and who have had bilateral salpingo-oophorectomy were shown to have improved overall and cancer-specific survival.
• In this specific group of BRCA mutation carriers, this study did not demonstrate an increased risk of mortality from cardiovascular disease, osteoporosis, or other causes associated with premature menopause from bilateral salpingo-oophorectomy.

In BRCA carriers, HRT after oophorectomy does not raise breast cancer risk

Rebbeck TR, Friebel T, Wagner T, et al; for the PROSE Study Group. Effect of short-term hormone replacement therapy on breast cancer risk reduction after bilateral prophylactic oophorectomy in BRCA1 and BRCA2 mutation carriers: the PROSE Study Group. J Clin Oncol. 2005;23:7804–7810.

In women with BRCA1 or BRCA2 mutations, the risk of ovarian cancer is staggering (20% to 40% for BRCA1 mutations, 15% to 25% for BRCA2 mutations), and prophylactic bilateral salpingo-oophorectomy is the only strategy proven to significantly reduce risk. A second important benefit for bilateral salpingo-oophorectomy among premenopausal mutation carriers is that the procedure decreases the risk of breast cancer by 50%. In women who have a lifetime risk of breast cancer that is as high as 80%, this benefit is extremely welcome. Current recommendations are for women with BRCA1 or BRCA2 mutations to undergo bilateral salpingo-oophorectomy at age 35 to 40, or when child-bearing is complete.

Yet, for many women in this age group, quality of life is substantially altered when premature menopause kicks in after the surgery. Most ObGyns feel comfortable giving short-term hormone replacement therapy (HRT) after prophylactic bilateral salpingo-oophorectomy to premenopausal women who do not have a history of breast cancer. However, until this study, no data were available that addressed the question of whether short-term HRT affects breast cancer risk.

In a prospective cohort of 462 women, of whom 155 underwent bilateral salpingo-oophorectomy, Rebbeck and colleagues evaluated the risk of developing breast cancer over an average of 3.6 years based on exposure to any type of HRT. In this multicenter study conducted at 13 different institutions in the United States and Europe, they found that women who underwent bilateral salpingo-oophorectomy for a BRCA1 or BRCA2 mutation were more likely to be older, have had children, and were more likely to use HRT. Compared with women who did not have bilateral salpingo-oophorectomy, women who had undergone the procedure and used any short-term HRT (including estrogen, progesterone, or a combination) still had a substantial decrease in breast cancer risk (HR=0.37; 95% CI, 0.14–0.96).

Practice recommendations

We can now reassure young women who must decide whether to undergo prophylactic bilateral salpingo-oophorectomy to reduce their staggering risks of breast and ovarian cancer: Short-term HRT to address the hot flashes, night sweats, and vaginal dryness associated with premature surgical menopause, first, is clinically reasonable and, second, will not substantially reduce the benefits of bilateral salpingo-oophorectomy for breast cancer risk.

ObGyns perform most of the screening for cancers of the ovary and breast. The first cancer is especially lethal, though rare, and the second is especially feared among women. This update reviews screening guidelines and recent studies that may affect how we detect and prevent ovarian and breast cancers.

Among the findings:

• In the only multicenter, prospective, randomized, controlled study to date to look at the use of CA-125 and transvaginal ultrasound screening in a low-risk population of postmenopausal women in the United States, researchers found no evidence to suggest a need to revise the present (1996) ovarian cancer screening guidelines of the US Preventive Services Task Force.
• Using a Markov decision-analysis model, investigators explored the health effects of prophylactic bilateral salpingo-oophorectomy in women at average risk of ovarian cancer undergoing hysterectomy. They found that removing the ovaries may decrease overall survival.
• Investigators found the opposite to be true in women with BRCA1 or BRCA2 mutations. Prophylactic bilateral salpingo-oophorectomy greatly reduced the overall mortality rate, as well as the risk of ovarian and breast cancer.
• In a prospective cohort study of BRCA mutation carriers with no history of breast cancer who underwent prophylactic oophorectomy, researchers found the short-term use of hormone replacement therapy to be safe, with no loss of protection against breast cancer.

No need to revise screening guidelines for ovarian cancer

Buys S, Partridge E, Greene M, et al; for the PLCO Project Team. Ovarian cancer screening in the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial: findings from the initial screen of a randomized trial. Am J Obstet Gynecol. 2005;193:1630–1639.

The need to identify a marker for the early detection of ovarian cancer is especially urgent, given that approximately 75% of women with the cancer present with late-stage disease. Because the disease is rare, finding a cost-effective screening test with good sensitivity and very high specificity (to decrease too many false-positive results) will be challenging.

So far, no prospective, randomized studies of any ovarian cancer screening modality have demonstrated a decrease in mortality—the gold standard of efficacy for any screening test. Therefore, the Prostate, Lung, Colorectal, and Ovarian (PLCO) cancer trial is a critical study—it is the only multicenter, prospective, randomized, controlled study in the United States to tackle the question of whether CA-125 and transvaginal ultrasonographic (US) screening will be effective in a low-risk population of postmenopausal women aged 55 to 74.

In this large study, 1 arm underwent ovarian cancer screening with both modalities and the other arm underwent no such screening.

This study reports on baseline, or ‘prevalent,’ cancers

This preliminary report does not comment on the efficacy of ovarian cancer screening; data on the effect of repeated annual screens on detection rates and mortality will become available over the next several years.

Rather, the purpose of this preliminary report was to detail the baseline ovarian cancer screening tests of the 39,115 women randomized to the intervention arm from November 15, 1993, to December 13, 2001. These results describe “prevalent” cancers—that is, cancers that are present on the first screen. The more important information about efficacy of screening will come over the next several years, as “incident” cancers develop.

Roughly 6% of women had at least 1 abnormal finding at baseline

Among 28,506 women with results for both baseline tests, 1,706 had at least 1 abnormal finding:

• 1,338 had an abnormal transvaginal US scan
• 402 had an abnormal level on the CA-125 test
• 34 had abnormalities in both tests
• 29 malignant neoplasms were identified in this population, 20 of them invasive.

When combined, CA-125 and transvaginal ultrasonography had good positive predictive value

In general, a positive predictive value (PPV) of more than 10% (ie, 10 surgeries to detect 1 cancer) is considered reasonable justification for a screening test. In the PLCO trial, the PPV was 4% for CA-125 alone (16 neoplasms in 402 positive screens), 1.6% for transvaginal ultrasonography alone (22 neoplasms in 1,338 positive screens), and 26.5% if both tests were abnormal (9 neoplasms in 34 positive screens).

When tumors of low malignant potential were excluded, the PPV was 3.7% for an abnormal CA-125, 1.0% for an abnormal transvaginal sonogram, and 23.5% if both tests were abnormal. A PPV of 23.5% for both tests is fairly good (ie, approximately 4 surgeries to detect 1 cancer). However, if only women in whom both screening tests were abnormal went to surgery, 12 of 20 invasive cancers would be missed.

 

Bottom line: Routine screening still not justified

Nothing in the findings reported here suggests that we need to revise the current (from 1996) ovarian cancer screening guidelines of the US Preventive Services Task Force,¹ which state that “routine screening for ovarian cancer by US, the measurement of serum tumor markers, or pelvic examination is not recommended.”

We will need to wait until the PLCO trial results come in to see the effect of repeated annual ovarian cancer screens on detection rates and mortality.

Consider ovarian conservation in hysterectomy for benign disease

Parker W, Broder MS, Liu Z, Shoupe D, Farquhar C, Berek JS. Ovarian conservation at the time of hysterectomy for benign disease. Obstet Gynecol. 2005;106:219–226.

As discussed earlier, we have no good screening tool for the early detection of ovarian cancer. Although rare, ovarian cancer is a lethal, scary disease, and most ObGyns prophylactically remove the ovaries at the time of hysterectomy in most postmenopausal and many perimenopausal women.

The downside to this strategy seems low among postmenopausal women, and the upside, in terms of not having to worry about ovarian cancer, seems high. The study by Parker and colleagues, while having definite limitations, asks us to question this routine practice pattern. The authors found that prophylactic oophorectomy may be associated with decreased overall survival.

Model used SEER data, Nurses’ Health Study to predict survival

Parker and colleagues used a Markov decision-analysis model (a hypothetical mathematical model that uses published data to create cohorts of patients to estimate risk of morbidity or mortality, or both, over time) to evaluate the risks and benefits of ovarian conservation at the time of hysterectomy for benign disease. Age-specific mortality estimates for ovarian cancer were based on Surveillance, Epidemiology and End Results (SEER) statistics.

For women at average risk of ovarian cancer, the probability of surviving to 80 years of age after hysterectomy between 50 and 54 years varied, and was 62.8% and 62.5% for ovarian conservation with and without estrogen therapy, respectively, compared with 62.2% and 53.9% for oophorectomy with and without estrogen therapy. The main reason that the model found decreased overall survival with prophylactic oophorectomy was an increase in coronary artery disease after oophorectomy—a finding that was based on data from the Nurses’ Health Study.

At the very least, think hard about the decision to remove the ovaries

This report estimated that about 300,000 prophylactic oophorectomies are carried out annually in the United States. Although this study has limitations, we believe it encourages debate and reexamination of the benefit of prophylactic oophorectomy for benign indications in young, low-risk patients.

The most important finding from the study is that oophorectomy conferred no survival advantage. Given the rarity of ovarian cancer among the general population, this effect is not that surprising.

For now, careful risk assessment remains a fundamental component of management, so that women who are at increased risk of ovarian cancer can undergo prophylactic salpingo-oophorectomy.

For low-risk women, who constitute the majority of patients, the data to support removing ovaries at the time of hysterectomy are less clear.

Make sure the patient understands the low risk of cancer and possible cardiac benefits of preservation

• Conduct a thorough discussion with the patient about the pros and cons of oophorectomy for benign disease.
• The risk of ovarian cancer in the general population is low; patients should understand that they may derive cardiac protection from postmenopausal ovarian function.

Do consider oophorectomy among carriers of a BRCA mutation

Domchek SM, Friebel TM, Neuhausen SL, et al. Mortality after bilateral salpingo-oophorectomy in BRCA1 and BRCA2 mutation carriers: a prospective cohort study. Lancet Oncol. 2006;7:223–229.

Women known to have BRCA1 or BRCA2 mutations are now managed by means of surveillance or with prophylactic bilateral salpingo-oophorectomy. Although bilateral salpingo-oophorectomy has been shown to reduce the risk of ovarian cancer by 90% and the risk of breast cancer by 50%, until this study few data shed light on the effect of the procedure on overall mortality among women with BRCA mutations.

This prospective cohort study identified 155 patients with BRCA1 or BRCA2 mutations who elected to undergo bilateral salpingo-oophorectomy and matched them by age with a control group of 271. All women were followed until death by any cause or by breast, ovarian, or primary peritoneal cancer. The women were followed for a mean of 3.1 years in the bilateral salpingo-oophorectomy group and 2.1 years in the control group.

Overall and cancer-specific survival improved with oophorectomy

Among BRCA mutation carriers, women who chose prophylactic bilateral salpingo-oophorectomy had improved overall and cancer-specific survival, compared with women who did not undergo the surgery.

 

In the analysis of the matched BRCA mutation carriers, women who chose to undergo the procedure had a decreased risk of overall mortality (hazard ratio [HR]=0.24; 95% confidence interval [CI], 0.08–0.71); they also had a decreased risk of mortality due to both breast cancer (HR=0.1; 95% CI, 0.02–0.71) and ovarian cancer (HR=0.05; 95% CI, 0.01–0.46).

Practice recommendations

Apparently, unlike women at average risk of ovarian cancer (for whom prophylactic oophorectomy in conjunction with hysterectomy for benign disease may be associated with decreased overall survival; see the review of the study by Parker and colleagues), women with BRCA mutations may benefit from oophorectomy.

Advantages of prophylactic bilateral salpingo-oophorectomy in this patient population should be discussed with potential surgical candidates, because:

• Women who have a BRCA mutation and who have had bilateral salpingo-oophorectomy were shown to have improved overall and cancer-specific survival.
• In this specific group of BRCA mutation carriers, this study did not demonstrate an increased risk of mortality from cardiovascular disease, osteoporosis, or other causes associated with premature menopause from bilateral salpingo-oophorectomy.

In BRCA carriers, HRT after oophorectomy does not raise breast cancer risk

Rebbeck TR, Friebel T, Wagner T, et al; for the PROSE Study Group. Effect of short-term hormone replacement therapy on breast cancer risk reduction after bilateral prophylactic oophorectomy in BRCA1 and BRCA2 mutation carriers: the PROSE Study Group. J Clin Oncol. 2005;23:7804–7810.

In women with BRCA1 or BRCA2 mutations, the risk of ovarian cancer is staggering (20% to 40% for BRCA1 mutations, 15% to 25% for BRCA2 mutations), and prophylactic bilateral salpingo-oophorectomy is the only strategy proven to significantly reduce risk. A second important benefit for bilateral salpingo-oophorectomy among premenopausal mutation carriers is that the procedure decreases the risk of breast cancer by 50%. In women who have a lifetime risk of breast cancer that is as high as 80%, this benefit is extremely welcome. Current recommendations are for women with BRCA1 or BRCA2 mutations to undergo bilateral salpingo-oophorectomy at age 35 to 40, or when child-bearing is complete.

Yet, for many women in this age group, quality of life is substantially altered when premature menopause kicks in after the surgery. Most ObGyns feel comfortable giving short-term hormone replacement therapy (HRT) after prophylactic bilateral salpingo-oophorectomy to premenopausal women who do not have a history of breast cancer. However, until this study, no data were available that addressed the question of whether short-term HRT affects breast cancer risk.

In a prospective cohort of 462 women, of whom 155 underwent bilateral salpingo-oophorectomy, Rebbeck and colleagues evaluated the risk of developing breast cancer over an average of 3.6 years based on exposure to any type of HRT. In this multicenter study conducted at 13 different institutions in the United States and Europe, they found that women who underwent bilateral salpingo-oophorectomy for a BRCA1 or BRCA2 mutation were more likely to be older, have had children, and were more likely to use HRT. Compared with women who did not have bilateral salpingo-oophorectomy, women who had undergone the procedure and used any short-term HRT (including estrogen, progesterone, or a combination) still had a substantial decrease in breast cancer risk (HR=0.37; 95% CI, 0.14–0.96).

Practice recommendations

We can now reassure young women who must decide whether to undergo prophylactic bilateral salpingo-oophorectomy to reduce their staggering risks of breast and ovarian cancer: Short-term HRT to address the hot flashes, night sweats, and vaginal dryness associated with premature surgical menopause, first, is clinically reasonable and, second, will not substantially reduce the benefits of bilateral salpingo-oophorectomy for breast cancer risk.

ObGyns perform most of the screening for cancers of the ovary and breast. The first cancer is especially lethal, though rare, and the second is especially feared among women. This update reviews screening guidelines and recent studies that may affect how we detect and prevent ovarian and breast cancers.

Among the findings:

• In the only multicenter, prospective, randomized, controlled study to date to look at the use of CA-125 and transvaginal ultrasound screening in a low-risk population of postmenopausal women in the United States, researchers found no evidence to suggest a need to revise the present (1996) ovarian cancer screening guidelines of the US Preventive Services Task Force.
• Using a Markov decision-analysis model, investigators explored the health effects of prophylactic bilateral salpingo-oophorectomy in women at average risk of ovarian cancer undergoing hysterectomy. They found that removing the ovaries may decrease overall survival.
• Investigators found the opposite to be true in women with BRCA1 or BRCA2 mutations. Prophylactic bilateral salpingo-oophorectomy greatly reduced the overall mortality rate, as well as the risk of ovarian and breast cancer.
• In a prospective cohort study of BRCA mutation carriers with no history of breast cancer who underwent prophylactic oophorectomy, researchers found the short-term use of hormone replacement therapy to be safe, with no loss of protection against breast cancer.

No need to revise screening guidelines for ovarian cancer

Buys S, Partridge E, Greene M, et al; for the PLCO Project Team. Ovarian cancer screening in the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial: findings from the initial screen of a randomized trial. Am J Obstet Gynecol. 2005;193:1630–1639.

The need to identify a marker for the early detection of ovarian cancer is especially urgent, given that approximately 75% of women with the cancer present with late-stage disease. Because the disease is rare, finding a cost-effective screening test with good sensitivity and very high specificity (to decrease too many false-positive results) will be challenging.

So far, no prospective, randomized studies of any ovarian cancer screening modality have demonstrated a decrease in mortality—the gold standard of efficacy for any screening test. Therefore, the Prostate, Lung, Colorectal, and Ovarian (PLCO) cancer trial is a critical study—it is the only multicenter, prospective, randomized, controlled study in the United States to tackle the question of whether CA-125 and transvaginal ultrasonographic (US) screening will be effective in a low-risk population of postmenopausal women aged 55 to 74.

In this large study, 1 arm underwent ovarian cancer screening with both modalities and the other arm underwent no such screening.

This study reports on baseline, or ‘prevalent,’ cancers

This preliminary report does not comment on the efficacy of ovarian cancer screening; data on the effect of repeated annual screens on detection rates and mortality will become available over the next several years.

Rather, the purpose of this preliminary report was to detail the baseline ovarian cancer screening tests of the 39,115 women randomized to the intervention arm from November 15, 1993, to December 13, 2001. These results describe “prevalent” cancers—that is, cancers that are present on the first screen. The more important information about efficacy of screening will come over the next several years, as “incident” cancers develop.

Roughly 6% of women had at least 1 abnormal finding at baseline

Among 28,506 women with results for both baseline tests, 1,706 had at least 1 abnormal finding:

• 1,338 had an abnormal transvaginal US scan
• 402 had an abnormal level on the CA-125 test
• 34 had abnormalities in both tests
• 29 malignant neoplasms were identified in this population, 20 of them invasive.

When combined, CA-125 and transvaginal ultrasonography had good positive predictive value

In general, a positive predictive value (PPV) of more than 10% (ie, 10 surgeries to detect 1 cancer) is considered reasonable justification for a screening test. In the PLCO trial, the PPV was 4% for CA-125 alone (16 neoplasms in 402 positive screens), 1.6% for transvaginal ultrasonography alone (22 neoplasms in 1,338 positive screens), and 26.5% if both tests were abnormal (9 neoplasms in 34 positive screens).

When tumors of low malignant potential were excluded, the PPV was 3.7% for an abnormal CA-125, 1.0% for an abnormal transvaginal sonogram, and 23.5% if both tests were abnormal. A PPV of 23.5% for both tests is fairly good (ie, approximately 4 surgeries to detect 1 cancer). However, if only women in whom both screening tests were abnormal went to surgery, 12 of 20 invasive cancers would be missed.

 

Bottom line: Routine screening still not justified

Nothing in the findings reported here suggests that we need to revise the current (from 1996) ovarian cancer screening guidelines of the US Preventive Services Task Force,¹ which state that “routine screening for ovarian cancer by US, the measurement of serum tumor markers, or pelvic examination is not recommended.”

We will need to wait until the PLCO trial results come in to see the effect of repeated annual ovarian cancer screens on detection rates and mortality.

Consider ovarian conservation in hysterectomy for benign disease

Parker W, Broder MS, Liu Z, Shoupe D, Farquhar C, Berek JS. Ovarian conservation at the time of hysterectomy for benign disease. Obstet Gynecol. 2005;106:219–226.

As discussed earlier, we have no good screening tool for the early detection of ovarian cancer. Although rare, ovarian cancer is a lethal, scary disease, and most ObGyns prophylactically remove the ovaries at the time of hysterectomy in most postmenopausal and many perimenopausal women.

The downside to this strategy seems low among postmenopausal women, and the upside, in terms of not having to worry about ovarian cancer, seems high. The study by Parker and colleagues, while having definite limitations, asks us to question this routine practice pattern. The authors found that prophylactic oophorectomy may be associated with decreased overall survival.

Model used SEER data, Nurses’ Health Study to predict survival

Parker and colleagues used a Markov decision-analysis model (a hypothetical mathematical model that uses published data to create cohorts of patients to estimate risk of morbidity or mortality, or both, over time) to evaluate the risks and benefits of ovarian conservation at the time of hysterectomy for benign disease. Age-specific mortality estimates for ovarian cancer were based on Surveillance, Epidemiology and End Results (SEER) statistics.

For women at average risk of ovarian cancer, the probability of surviving to 80 years of age after hysterectomy between 50 and 54 years varied, and was 62.8% and 62.5% for ovarian conservation with and without estrogen therapy, respectively, compared with 62.2% and 53.9% for oophorectomy with and without estrogen therapy. The main reason that the model found decreased overall survival with prophylactic oophorectomy was an increase in coronary artery disease after oophorectomy—a finding that was based on data from the Nurses’ Health Study.

At the very least, think hard about the decision to remove the ovaries

This report estimated that about 300,000 prophylactic oophorectomies are carried out annually in the United States. Although this study has limitations, we believe it encourages debate and reexamination of the benefit of prophylactic oophorectomy for benign indications in young, low-risk patients.

The most important finding from the study is that oophorectomy conferred no survival advantage. Given the rarity of ovarian cancer among the general population, this effect is not that surprising.

For now, careful risk assessment remains a fundamental component of management, so that women who are at increased risk of ovarian cancer can undergo prophylactic salpingo-oophorectomy.

For low-risk women, who constitute the majority of patients, the data to support removing ovaries at the time of hysterectomy are less clear.

Make sure the patient understands the low risk of cancer and possible cardiac benefits of preservation

• Conduct a thorough discussion with the patient about the pros and cons of oophorectomy for benign disease.
• The risk of ovarian cancer in the general population is low; patients should understand that they may derive cardiac protection from postmenopausal ovarian function.

Do consider oophorectomy among carriers of a BRCA mutation

Domchek SM, Friebel TM, Neuhausen SL, et al. Mortality after bilateral salpingo-oophorectomy in BRCA1 and BRCA2 mutation carriers: a prospective cohort study. Lancet Oncol. 2006;7:223–229.

Women known to have BRCA1 or BRCA2 mutations are now managed by means of surveillance or with prophylactic bilateral salpingo-oophorectomy. Although bilateral salpingo-oophorectomy has been shown to reduce the risk of ovarian cancer by 90% and the risk of breast cancer by 50%, until this study few data shed light on the effect of the procedure on overall mortality among women with BRCA mutations.

This prospective cohort study identified 155 patients with BRCA1 or BRCA2 mutations who elected to undergo bilateral salpingo-oophorectomy and matched them by age with a control group of 271. All women were followed until death by any cause or by breast, ovarian, or primary peritoneal cancer. The women were followed for a mean of 3.1 years in the bilateral salpingo-oophorectomy group and 2.1 years in the control group.

Overall and cancer-specific survival improved with oophorectomy

Among BRCA mutation carriers, women who chose prophylactic bilateral salpingo-oophorectomy had improved overall and cancer-specific survival, compared with women who did not undergo the surgery.

 

In the analysis of the matched BRCA mutation carriers, women who chose to undergo the procedure had a decreased risk of overall mortality (hazard ratio [HR]=0.24; 95% confidence interval [CI], 0.08–0.71); they also had a decreased risk of mortality due to both breast cancer (HR=0.1; 95% CI, 0.02–0.71) and ovarian cancer (HR=0.05; 95% CI, 0.01–0.46).

Practice recommendations

Apparently, unlike women at average risk of ovarian cancer (for whom prophylactic oophorectomy in conjunction with hysterectomy for benign disease may be associated with decreased overall survival; see the review of the study by Parker and colleagues), women with BRCA mutations may benefit from oophorectomy.

Advantages of prophylactic bilateral salpingo-oophorectomy in this patient population should be discussed with potential surgical candidates, because:

• Women who have a BRCA mutation and who have had bilateral salpingo-oophorectomy were shown to have improved overall and cancer-specific survival.
• In this specific group of BRCA mutation carriers, this study did not demonstrate an increased risk of mortality from cardiovascular disease, osteoporosis, or other causes associated with premature menopause from bilateral salpingo-oophorectomy.

In BRCA carriers, HRT after oophorectomy does not raise breast cancer risk

Rebbeck TR, Friebel T, Wagner T, et al; for the PROSE Study Group. Effect of short-term hormone replacement therapy on breast cancer risk reduction after bilateral prophylactic oophorectomy in BRCA1 and BRCA2 mutation carriers: the PROSE Study Group. J Clin Oncol. 2005;23:7804–7810.

In women with BRCA1 or BRCA2 mutations, the risk of ovarian cancer is staggering (20% to 40% for BRCA1 mutations, 15% to 25% for BRCA2 mutations), and prophylactic bilateral salpingo-oophorectomy is the only strategy proven to significantly reduce risk. A second important benefit for bilateral salpingo-oophorectomy among premenopausal mutation carriers is that the procedure decreases the risk of breast cancer by 50%. In women who have a lifetime risk of breast cancer that is as high as 80%, this benefit is extremely welcome. Current recommendations are for women with BRCA1 or BRCA2 mutations to undergo bilateral salpingo-oophorectomy at age 35 to 40, or when child-bearing is complete.

Yet, for many women in this age group, quality of life is substantially altered when premature menopause kicks in after the surgery. Most ObGyns feel comfortable giving short-term hormone replacement therapy (HRT) after prophylactic bilateral salpingo-oophorectomy to premenopausal women who do not have a history of breast cancer. However, until this study, no data were available that addressed the question of whether short-term HRT affects breast cancer risk.

In a prospective cohort of 462 women, of whom 155 underwent bilateral salpingo-oophorectomy, Rebbeck and colleagues evaluated the risk of developing breast cancer over an average of 3.6 years based on exposure to any type of HRT. In this multicenter study conducted at 13 different institutions in the United States and Europe, they found that women who underwent bilateral salpingo-oophorectomy for a BRCA1 or BRCA2 mutation were more likely to be older, have had children, and were more likely to use HRT. Compared with women who did not have bilateral salpingo-oophorectomy, women who had undergone the procedure and used any short-term HRT (including estrogen, progesterone, or a combination) still had a substantial decrease in breast cancer risk (HR=0.37; 95% CI, 0.14–0.96).

Practice recommendations

We can now reassure young women who must decide whether to undergo prophylactic bilateral salpingo-oophorectomy to reduce their staggering risks of breast and ovarian cancer: Short-term HRT to address the hot flashes, night sweats, and vaginal dryness associated with premature surgical menopause, first, is clinically reasonable and, second, will not substantially reduce the benefits of bilateral salpingo-oophorectomy for breast cancer risk.

Multiple gestations are far more common today than they once were—up 70% since 1980.¹ Today, every 1,000 live births include 32.3 sets of twins, a 2% increase in the rate of twin births since 2004. Why this phenomenon is occurring is not entirely understood but, certainly, the trend toward older maternal age and the emergence of assisted reproduction are both part of the explanation.

Multiple gestations are of particular concern to obstetricians because, even though they remain relatively rare, they are responsible for a significant percentage of perinatal morbidity and mortality.

The difficulties that twins encounter are often associated with preterm birth and occur most often in identical twins developing within a single gestational sac. Those difficulties include malformation, chromosomal abnormalities, learning disability, behavioral problems, chronic lung disease, neuromuscular developmental delay, cerebral palsy, and stillbirth. Women pregnant with twins are also at heightened risk, particularly of gestational hypertension, preeclampsia, and gestational diabetes.²

Your task is to manage these risks so that the outlook for mother and infant is as favorable as possible.

Determining chorionicity in the first trimester

Ultrasonographic determination of chorionicity should be the first step in the management of a twin gestation. The determination should be made as early as possible in the pregnancy because it has an immediate impact on counseling, risk of miscarriage, and efficacy of noninvasive screening. (See “Is there 1 sac, or more? Key to predicting risk.”)

The accuracy of ultrasonography (US) in determining chorionicity depends on gestational age. US predictors of dichorionicity include:

• gender discordance
• separate placentas
• the so-called twin-peak sign (also called the lambda sign) (FIGURE 1), in which the placenta appears to extend a short distance between the gestational sacs; compare this with FIGURE 2, showing monochorionic twins with the absence of an intervening placenta
• an intertwin membrane thicker than 1.5 mm to 2.0 mm.

US examination can accurately identify chorionicity at 10 to 14 weeks’ gestation, with overall sensitivity that is reported to be as high as 100%.^3-5

FIGURE 1

The twin-peak sign on US

This dichorionic–diamnionic twin gestation demonstrates the so-called twin peak, or lambda, sign (arrow), in which the placenta appears to extend a short distance between the gestational sacs.

FIGURE 2

Monochorionic–diamnionic twin gestation

This US scan of a monochorionic twin gestation reveals the absence of an intervening placenta

What can go wrong

Monochorionic twins

Twins who share a gestational sac are more likely than 2-sac twins to suffer spontaneous loss, congenital anomalies, growth restriction and discordancy, preterm delivery, and neurologic morbidity.

Spontaneous loss. In 1 comparative series, the risk of pregnancy loss at less than 24 weeks’ gestation was 12.2% for monochorionic twins, compared with 1.8% for dichorionic twins.⁶ Spontaneously conceived monochorionic twins may have the highest risk of loss.⁷ However, monochorionic twins occur more often in conceptions achieved by assisted reproductive technology—at a rate 3 to 10 times higher than the background rate of monochorionic twinning.⁸

KEY POINTS

Congenital anomalies. These occur 2 to 3 times as often in monochorionic twins, and have been reported in as many as 10% of such pregnancies. Reported anomalies include midline defects, cloacal abnormalities, neural tube defects, ventral wall defects, craniofacial abnormalities, conjoined twins, and acardiac twins.^9-11 In light of these risks, a detailed anatomic survey is suggested for all twins.

 

Heart defects. The incidence of congenital heart defects is 4 times greater in monochorionic twins, even in the absence of twin-to-twin transfusion syndrome (TTTS).¹² Cardiac malformations may occur secondary to abnormal lateralization during embryogenesis or result from an abnormal vascular distribution in the shared placenta.^9,10 The presence of abnormal vascular communications may also cause limb reduction defects and the rare acardiac twin.

Long-term neurologic morbidity. In one series, the incidence of cerebral palsy was 8%, compared with 1% among dichorionic twins. In twins followed to 2 years of age, rates of minor neurologic morbidity were 15% in monochorionic twins and 3% in dichorionic twins. The overall rate of neurologic disorders in monochorionic twins was 23%, regardless of fetal weight.¹³ At 4 years, long-term neurologic morbidity was particularly high in single survivors of a monochorionic pair; the incidence of cerebral palsy has been reported to be as high as 50% in single survivors, compared with 14.3% in cases in which both twins survived.¹⁴

Twin-to-twin transfusion syndrome. In this condition, abnormal vascular connections arise in the shared placenta, allowing blood to be shunted from one fetus to the other. The syndrome is unique to monochorionic gestations and occurs in 15% to 20% of cases.¹⁵ A significant percentage of neurologic morbidity is probably the result of TTTS. To evaluate for TTTS, include a detailed anatomic survey and serial US every 2 weeks beginning in the second trimester as part of the surveillance of monochorionic twin gestations. (See “TTTS: Diagnosis, staging, treatment.”)

Down syndrome and other chromosomal abnormalities

Estimating odds

Assessing the likelihood of a chromosomal abnormality (aneuploidy) in a multiple gestation is complicated by differences in twinning mechanisms (chorionicity versus zygosity) and by the increasing rate of dizygotic twinning with advancing maternal age. The risk is greater in dizygotic twin gestations than in age-matched singleton gestations. The definition of advanced maternal age (AMA) in a twin pregnancy has ranged from 31 to 33 years of age in reports in the literature.^2,16,17

The probability that a twin gestation contains a fetus with a chromosomal abnormality is directly related to zygosity. Each twin in a dizygotic gestation carries an independent risk, so the composite risk for the pregnancy is a summation of the independent risk for each fetus. For monozygotic twins, the risk is similar to the age-related risk in a singleton gestation. Presumptions about zygosity are based on chorionicity: Almost all (90%) dichorionic twins are dizygotic and all monochorionic twins are monozygotic.

What is the utility of noninvasive screening?

Multiple gestations can be screened for aneuploidy using maternal age, maternal serum markers, and nuchal translucency (NT) on US, or combinations of these assessments.

When first-trimester serum markers (free β-human chorionic gonadotropin and pregnancy-associated plasma protein A [PAPPA]) are combined with NT and maternal age, a pregnancy-specific risk can be calculated that includes the individual contribution of each fetus, thus yielding an improved detection rate. In monochorionic twins, the NTs are averaged to calculate a single risk for the entire pregnancy. In dichorionic twins, the risk for each fetus is calculated independently and then summed to establish a pregnancy-specific risk. The combined test has a reported detection rate of 84% for monochorionic twins and 70% for dichorionic twins, compared with detection rates of 85% to 87% for singletons at a 5% false-positive rate.^18,19 The integrated test (combined test plus measurement of second-trimester serum analytes) has a 93% detection rate for monochorionic twins and a 78% detection rate for dichorionic twins, compared with 95% to 96% for singletons at the same 5% false-positive rate.^18,19 Second-trimester screening has a lower detection rate in both singleton and twin gestations.

 

Prenatal diagnosis

Given the lower detection rate of aneuploidy in twin gestations and the associated increase in aneuploidy with advancing maternal age, many patients choose to undergo prenatal diagnosis rather than relying on screening. On the basis of maternal age alone, invasive prenatal diagnosis can be offered to women who will be 31 years or older at their estimated due date.

Available diagnostic options include chorionic villus sampling (CVS) or amniocentesis. CVS is performed at an earlier gestational age (10 to 13 weeks) than amniocentesis (15 to 20 weeks). Multiples pose specific technical considerations for either procedure, and accurate fetal mapping is essential. Successful sampling with CVS can be performed in more than 99% of cases; the rate of cross-contamination is less than 1%.

Is there a risk of miscarriage?

In counseling patients about the risk of fetal death that CVS or amniocentesis may entail, the place to begin is the background loss rate, which is greater in twin than in single gestations. The reported background loss rate of twins at 24 weeks’ gestation or less ranges from 5.8% to 7.2%.^20,21 In women of advanced maternal age (35 years and older), a background rate as high as 17.6% has been described.²¹ Once parents are aware of this, they have a context for weighing the risk of miscarriage that prenatal testing may hold.

 

The twin loss rate following amniocentesis has been evaluated in several studies. (See “Integrating evidence and experience: Does invasive prenatal testing raise the risk of miscarriage in a twin gestation?”)

A greatly elevated risk of preterm birth

Multiple gestations are at extremely high risk for premature delivery, and—like all premature newborns—these infants are at risk for a wide range of disabilities. Risk factors for premature delivery include history of second trimester pregnancy loss, preterm birth at less than 35 weeks’ gestation, more than 2 previous curettage procedures, cone biopsy, müllerian anomaly, and diethylstilbestrol exposure. Unfortunately, current yardsticks for predicting premature delivery in multiple pregnancy have serious limitations, and available interventions have not been particularly successful.

Predictors

Measurement of cervical length has been evaluated as a predictor of preterm delivery in a number of twin studies that were looking for a cutoff point that can predict which twins are at greatest risk. No such cutoff has been found.^22-25

In general, studies demonstrate a low risk of preterm delivery for women who have a cervical length measurement of more than 35 mm at 24 to 26 weeks. A shorter cervical length correlates with premature delivery, but specific cutoffs have proved not to be sensitive predictors.

In the largest published series, To and colleagues evaluated 1,163 sets of twins undergoing routine care with cervical length assessments at 22 to 24 weeks’ gestation. They demonstrated a direct correlation between cervical length and preterm delivery, but were unable to define a cutoff sufficiently sensitive to be useful.²⁵ A shortened cervix may be predictive of prematurity in general, but it does not allow the obstetrician to predict with certainty which mothers will give birth prematurely or how long a particular mother will carry.

Fetal fibronectin. The presence of fetal fibronectin (ffN) in cervicovaginal secretions is widely used as an adjunct to other potential predictors of preterm delivery. In a multistudy review that included symptomatic women, a negative ffN had a 99% negative predictive value but a poor positive predictive value (13% to 30%) for delivery within 7 to 10 days.²⁶

Use of ffN in conjunction with cervical length has also been investigated in twin gestations. Although the negative predictive value of ffN remained high, the addition of ffN to cervical length assessment did not improve the positive predictive value of cervical length alone.^27,28

Interventions

Cerclage is often used in high-risk singleton pregnancies in which a shortened cervix is seen on a sonogram. The utility of cerclage in twins is less clear. Randomized controlled trials comparing women at risk of premature delivery treated with cerclage and controls not considered at risk found no difference in the rate of premature delivery in the 2 groups.^29,30 Meta-analysis of 4 randomized controlled trials also found no benefit and, in fact, detected a possibility of actual harm. Cerclage twins were more likely to deliver early (at less than 35 weeks’ gestation) and had a 2.6 relative risk of perinatal mortality. The differences found in the meta-analysis were not statistically significant, however, and the overall sample size was small (n=48).³¹

The best available data seem to show that cerclage based on US indications of cervical shortening is not beneficial and may even be associated with worse outcome.

17-Hydroxyprogesterone caproate (17P) has been found to decrease the rate of recurrent preterm birth in singleton gestations by almost 35%.³² Although twins are at increased risk of preterm birth, the use of 17P has not, however, been shown to be of benefit.³³

Bed rest. A Cochrane Database review of 6 randomized controlled trials compared 1) patients with a multiple gestation who were offered bed rest in the hospital with 2) patients hospitalized for complications of pregnancy. The review found that bed rest did not reduce the risk of preterm birth or of perinatal mortality in the routinely hospitalized women. There was, however, a tendency to a decreased number of low-birth-weight infants born to women given bed rest.³⁴

INTEGRATING EVIDENCE AND EXPERIENCE

 

Some parents elect fetal reduction

Given the high level of risk in a multiple pregnancy, reducing the number of fetuses is an option that some patients choose. A recent trend toward reduction to singleton pregnancy seems to be related to:

• increasing maternal age
• single parenthood
• financial considerations
• the increased medical risk to mother and fetuses associated with twins.³⁵

Evans and colleagues found that, although the overall rate of reduction from twins to a singleton was 3%, the percentage (76%) of women older than 35 years who opted for such a reduction was disproportionately high.³⁶

In a series of 1,000 cases of multifetal pregnancy reduction, Stone et al found that the pregnancy loss rate was lowest (2.5%) when there was reduction to a singleton gestation.³⁷ A comparative analysis of 2,000 cases of multifetal pregnancy reduction presented at the 2006 meeting of the Society for Maternal–Fetal Medicine found that the percentage of twin gestations undergoing reduction to a singleton has increased from 4% to 15.6% between 1999 and 2006, with an increase in the overall incidence of reduction to a singleton from 11.8% to 31.8%.³⁸

You should discuss pregnancy reduction with patients at high risk of pregnancy-associated complications such as cervical incompetence, preterm delivery, severe maternal cardiac disease, hypertension, diabetes, and uterine anomalies, as well as with patients who are carrying higher-order multiple gestations, in which the fetuses are at risk of problems.

Wrap-up: The tasks facing you in a multiple gestation

Start your management of a multiple gestation by taking the essential first step of determining the chorionicity of the fetuses.

Once you have done that, explain the risks of twin pregnancy and the particular risks of a single-sac pregnancy. Parents will want to know their risk of having a child with an anomaly; pay particular attention to the likelihood of a Down syndrome child. Noninvasive screening for Down syndrome and other chromosomal anomalies may be sufficient, but an older mother may prefer more definitive answers from CVS or amniocentesis.

You must prepare parents of a twin gestation for the risk of premature delivery. Cervical length assessment in the second trimester to early-third trimester may provide some indications of what is to happen, but the predictive value of this procedure is limited, and a shortened cervical length should be interpreted with caution.

In a monochorionic pregnancy, fetal growth should be assessed at regular intervals to evaluate for possible growth restriction or TTTS. If you detect evidence of abnormal fetal growth or amniotic fluid, US surveillance is indicated. Routine antepartum US surveillance of twins is not, however, recommended.²

The authors report no financial relationships relevant to this article.

Multiple gestations are far more common today than they once were—up 70% since 1980.¹ Today, every 1,000 live births include 32.3 sets of twins, a 2% increase in the rate of twin births since 2004. Why this phenomenon is occurring is not entirely understood but, certainly, the trend toward older maternal age and the emergence of assisted reproduction are both part of the explanation.

Multiple gestations are of particular concern to obstetricians because, even though they remain relatively rare, they are responsible for a significant percentage of perinatal morbidity and mortality.

The difficulties that twins encounter are often associated with preterm birth and occur most often in identical twins developing within a single gestational sac. Those difficulties include malformation, chromosomal abnormalities, learning disability, behavioral problems, chronic lung disease, neuromuscular developmental delay, cerebral palsy, and stillbirth. Women pregnant with twins are also at heightened risk, particularly of gestational hypertension, preeclampsia, and gestational diabetes.²

Your task is to manage these risks so that the outlook for mother and infant is as favorable as possible.

Determining chorionicity in the first trimester

Ultrasonographic determination of chorionicity should be the first step in the management of a twin gestation. The determination should be made as early as possible in the pregnancy because it has an immediate impact on counseling, risk of miscarriage, and efficacy of noninvasive screening. (See “Is there 1 sac, or more? Key to predicting risk.”)

The accuracy of ultrasonography (US) in determining chorionicity depends on gestational age. US predictors of dichorionicity include:

• gender discordance
• separate placentas
• the so-called twin-peak sign (also called the lambda sign) (FIGURE 1), in which the placenta appears to extend a short distance between the gestational sacs; compare this with FIGURE 2, showing monochorionic twins with the absence of an intervening placenta
• an intertwin membrane thicker than 1.5 mm to 2.0 mm.

US examination can accurately identify chorionicity at 10 to 14 weeks’ gestation, with overall sensitivity that is reported to be as high as 100%.^3-5

FIGURE 1

The twin-peak sign on US

This dichorionic–diamnionic twin gestation demonstrates the so-called twin peak, or lambda, sign (arrow), in which the placenta appears to extend a short distance between the gestational sacs.

FIGURE 2

Monochorionic–diamnionic twin gestation

This US scan of a monochorionic twin gestation reveals the absence of an intervening placenta

What can go wrong

Monochorionic twins

Twins who share a gestational sac are more likely than 2-sac twins to suffer spontaneous loss, congenital anomalies, growth restriction and discordancy, preterm delivery, and neurologic morbidity.

Spontaneous loss. In 1 comparative series, the risk of pregnancy loss at less than 24 weeks’ gestation was 12.2% for monochorionic twins, compared with 1.8% for dichorionic twins.⁶ Spontaneously conceived monochorionic twins may have the highest risk of loss.⁷ However, monochorionic twins occur more often in conceptions achieved by assisted reproductive technology—at a rate 3 to 10 times higher than the background rate of monochorionic twinning.⁸

KEY POINTS

Congenital anomalies. These occur 2 to 3 times as often in monochorionic twins, and have been reported in as many as 10% of such pregnancies. Reported anomalies include midline defects, cloacal abnormalities, neural tube defects, ventral wall defects, craniofacial abnormalities, conjoined twins, and acardiac twins.^9-11 In light of these risks, a detailed anatomic survey is suggested for all twins.

 

Heart defects. The incidence of congenital heart defects is 4 times greater in monochorionic twins, even in the absence of twin-to-twin transfusion syndrome (TTTS).¹² Cardiac malformations may occur secondary to abnormal lateralization during embryogenesis or result from an abnormal vascular distribution in the shared placenta.^9,10 The presence of abnormal vascular communications may also cause limb reduction defects and the rare acardiac twin.

Long-term neurologic morbidity. In one series, the incidence of cerebral palsy was 8%, compared with 1% among dichorionic twins. In twins followed to 2 years of age, rates of minor neurologic morbidity were 15% in monochorionic twins and 3% in dichorionic twins. The overall rate of neurologic disorders in monochorionic twins was 23%, regardless of fetal weight.¹³ At 4 years, long-term neurologic morbidity was particularly high in single survivors of a monochorionic pair; the incidence of cerebral palsy has been reported to be as high as 50% in single survivors, compared with 14.3% in cases in which both twins survived.¹⁴

Twin-to-twin transfusion syndrome. In this condition, abnormal vascular connections arise in the shared placenta, allowing blood to be shunted from one fetus to the other. The syndrome is unique to monochorionic gestations and occurs in 15% to 20% of cases.¹⁵ A significant percentage of neurologic morbidity is probably the result of TTTS. To evaluate for TTTS, include a detailed anatomic survey and serial US every 2 weeks beginning in the second trimester as part of the surveillance of monochorionic twin gestations. (See “TTTS: Diagnosis, staging, treatment.”)

Down syndrome and other chromosomal abnormalities

Estimating odds

Assessing the likelihood of a chromosomal abnormality (aneuploidy) in a multiple gestation is complicated by differences in twinning mechanisms (chorionicity versus zygosity) and by the increasing rate of dizygotic twinning with advancing maternal age. The risk is greater in dizygotic twin gestations than in age-matched singleton gestations. The definition of advanced maternal age (AMA) in a twin pregnancy has ranged from 31 to 33 years of age in reports in the literature.^2,16,17

The probability that a twin gestation contains a fetus with a chromosomal abnormality is directly related to zygosity. Each twin in a dizygotic gestation carries an independent risk, so the composite risk for the pregnancy is a summation of the independent risk for each fetus. For monozygotic twins, the risk is similar to the age-related risk in a singleton gestation. Presumptions about zygosity are based on chorionicity: Almost all (90%) dichorionic twins are dizygotic and all monochorionic twins are monozygotic.

What is the utility of noninvasive screening?

Multiple gestations can be screened for aneuploidy using maternal age, maternal serum markers, and nuchal translucency (NT) on US, or combinations of these assessments.

When first-trimester serum markers (free β-human chorionic gonadotropin and pregnancy-associated plasma protein A [PAPPA]) are combined with NT and maternal age, a pregnancy-specific risk can be calculated that includes the individual contribution of each fetus, thus yielding an improved detection rate. In monochorionic twins, the NTs are averaged to calculate a single risk for the entire pregnancy. In dichorionic twins, the risk for each fetus is calculated independently and then summed to establish a pregnancy-specific risk. The combined test has a reported detection rate of 84% for monochorionic twins and 70% for dichorionic twins, compared with detection rates of 85% to 87% for singletons at a 5% false-positive rate.^18,19 The integrated test (combined test plus measurement of second-trimester serum analytes) has a 93% detection rate for monochorionic twins and a 78% detection rate for dichorionic twins, compared with 95% to 96% for singletons at the same 5% false-positive rate.^18,19 Second-trimester screening has a lower detection rate in both singleton and twin gestations.

 

Prenatal diagnosis

Given the lower detection rate of aneuploidy in twin gestations and the associated increase in aneuploidy with advancing maternal age, many patients choose to undergo prenatal diagnosis rather than relying on screening. On the basis of maternal age alone, invasive prenatal diagnosis can be offered to women who will be 31 years or older at their estimated due date.

Available diagnostic options include chorionic villus sampling (CVS) or amniocentesis. CVS is performed at an earlier gestational age (10 to 13 weeks) than amniocentesis (15 to 20 weeks). Multiples pose specific technical considerations for either procedure, and accurate fetal mapping is essential. Successful sampling with CVS can be performed in more than 99% of cases; the rate of cross-contamination is less than 1%.

Is there a risk of miscarriage?

In counseling patients about the risk of fetal death that CVS or amniocentesis may entail, the place to begin is the background loss rate, which is greater in twin than in single gestations. The reported background loss rate of twins at 24 weeks’ gestation or less ranges from 5.8% to 7.2%.^20,21 In women of advanced maternal age (35 years and older), a background rate as high as 17.6% has been described.²¹ Once parents are aware of this, they have a context for weighing the risk of miscarriage that prenatal testing may hold.

 

The twin loss rate following amniocentesis has been evaluated in several studies. (See “Integrating evidence and experience: Does invasive prenatal testing raise the risk of miscarriage in a twin gestation?”)

A greatly elevated risk of preterm birth

Multiple gestations are at extremely high risk for premature delivery, and—like all premature newborns—these infants are at risk for a wide range of disabilities. Risk factors for premature delivery include history of second trimester pregnancy loss, preterm birth at less than 35 weeks’ gestation, more than 2 previous curettage procedures, cone biopsy, müllerian anomaly, and diethylstilbestrol exposure. Unfortunately, current yardsticks for predicting premature delivery in multiple pregnancy have serious limitations, and available interventions have not been particularly successful.

Predictors

Measurement of cervical length has been evaluated as a predictor of preterm delivery in a number of twin studies that were looking for a cutoff point that can predict which twins are at greatest risk. No such cutoff has been found.^22-25

In general, studies demonstrate a low risk of preterm delivery for women who have a cervical length measurement of more than 35 mm at 24 to 26 weeks. A shorter cervical length correlates with premature delivery, but specific cutoffs have proved not to be sensitive predictors.

In the largest published series, To and colleagues evaluated 1,163 sets of twins undergoing routine care with cervical length assessments at 22 to 24 weeks’ gestation. They demonstrated a direct correlation between cervical length and preterm delivery, but were unable to define a cutoff sufficiently sensitive to be useful.²⁵ A shortened cervix may be predictive of prematurity in general, but it does not allow the obstetrician to predict with certainty which mothers will give birth prematurely or how long a particular mother will carry.

Fetal fibronectin. The presence of fetal fibronectin (ffN) in cervicovaginal secretions is widely used as an adjunct to other potential predictors of preterm delivery. In a multistudy review that included symptomatic women, a negative ffN had a 99% negative predictive value but a poor positive predictive value (13% to 30%) for delivery within 7 to 10 days.²⁶

Use of ffN in conjunction with cervical length has also been investigated in twin gestations. Although the negative predictive value of ffN remained high, the addition of ffN to cervical length assessment did not improve the positive predictive value of cervical length alone.^27,28

Interventions

Cerclage is often used in high-risk singleton pregnancies in which a shortened cervix is seen on a sonogram. The utility of cerclage in twins is less clear. Randomized controlled trials comparing women at risk of premature delivery treated with cerclage and controls not considered at risk found no difference in the rate of premature delivery in the 2 groups.^29,30 Meta-analysis of 4 randomized controlled trials also found no benefit and, in fact, detected a possibility of actual harm. Cerclage twins were more likely to deliver early (at less than 35 weeks’ gestation) and had a 2.6 relative risk of perinatal mortality. The differences found in the meta-analysis were not statistically significant, however, and the overall sample size was small (n=48).³¹

The best available data seem to show that cerclage based on US indications of cervical shortening is not beneficial and may even be associated with worse outcome.

17-Hydroxyprogesterone caproate (17P) has been found to decrease the rate of recurrent preterm birth in singleton gestations by almost 35%.³² Although twins are at increased risk of preterm birth, the use of 17P has not, however, been shown to be of benefit.³³

Bed rest. A Cochrane Database review of 6 randomized controlled trials compared 1) patients with a multiple gestation who were offered bed rest in the hospital with 2) patients hospitalized for complications of pregnancy. The review found that bed rest did not reduce the risk of preterm birth or of perinatal mortality in the routinely hospitalized women. There was, however, a tendency to a decreased number of low-birth-weight infants born to women given bed rest.³⁴

INTEGRATING EVIDENCE AND EXPERIENCE

 

Some parents elect fetal reduction

Given the high level of risk in a multiple pregnancy, reducing the number of fetuses is an option that some patients choose. A recent trend toward reduction to singleton pregnancy seems to be related to:

• increasing maternal age
• single parenthood
• financial considerations
• the increased medical risk to mother and fetuses associated with twins.³⁵

Evans and colleagues found that, although the overall rate of reduction from twins to a singleton was 3%, the percentage (76%) of women older than 35 years who opted for such a reduction was disproportionately high.³⁶

In a series of 1,000 cases of multifetal pregnancy reduction, Stone et al found that the pregnancy loss rate was lowest (2.5%) when there was reduction to a singleton gestation.³⁷ A comparative analysis of 2,000 cases of multifetal pregnancy reduction presented at the 2006 meeting of the Society for Maternal–Fetal Medicine found that the percentage of twin gestations undergoing reduction to a singleton has increased from 4% to 15.6% between 1999 and 2006, with an increase in the overall incidence of reduction to a singleton from 11.8% to 31.8%.³⁸

You should discuss pregnancy reduction with patients at high risk of pregnancy-associated complications such as cervical incompetence, preterm delivery, severe maternal cardiac disease, hypertension, diabetes, and uterine anomalies, as well as with patients who are carrying higher-order multiple gestations, in which the fetuses are at risk of problems.

Wrap-up: The tasks facing you in a multiple gestation

Start your management of a multiple gestation by taking the essential first step of determining the chorionicity of the fetuses.

Once you have done that, explain the risks of twin pregnancy and the particular risks of a single-sac pregnancy. Parents will want to know their risk of having a child with an anomaly; pay particular attention to the likelihood of a Down syndrome child. Noninvasive screening for Down syndrome and other chromosomal anomalies may be sufficient, but an older mother may prefer more definitive answers from CVS or amniocentesis.

You must prepare parents of a twin gestation for the risk of premature delivery. Cervical length assessment in the second trimester to early-third trimester may provide some indications of what is to happen, but the predictive value of this procedure is limited, and a shortened cervical length should be interpreted with caution.

In a monochorionic pregnancy, fetal growth should be assessed at regular intervals to evaluate for possible growth restriction or TTTS. If you detect evidence of abnormal fetal growth or amniotic fluid, US surveillance is indicated. Routine antepartum US surveillance of twins is not, however, recommended.²

The authors report no financial relationships relevant to this article.

Multiple gestations are far more common today than they once were—up 70% since 1980.¹ Today, every 1,000 live births include 32.3 sets of twins, a 2% increase in the rate of twin births since 2004. Why this phenomenon is occurring is not entirely understood but, certainly, the trend toward older maternal age and the emergence of assisted reproduction are both part of the explanation.

Multiple gestations are of particular concern to obstetricians because, even though they remain relatively rare, they are responsible for a significant percentage of perinatal morbidity and mortality.

The difficulties that twins encounter are often associated with preterm birth and occur most often in identical twins developing within a single gestational sac. Those difficulties include malformation, chromosomal abnormalities, learning disability, behavioral problems, chronic lung disease, neuromuscular developmental delay, cerebral palsy, and stillbirth. Women pregnant with twins are also at heightened risk, particularly of gestational hypertension, preeclampsia, and gestational diabetes.²

Your task is to manage these risks so that the outlook for mother and infant is as favorable as possible.

Determining chorionicity in the first trimester

Ultrasonographic determination of chorionicity should be the first step in the management of a twin gestation. The determination should be made as early as possible in the pregnancy because it has an immediate impact on counseling, risk of miscarriage, and efficacy of noninvasive screening. (See “Is there 1 sac, or more? Key to predicting risk.”)

The accuracy of ultrasonography (US) in determining chorionicity depends on gestational age. US predictors of dichorionicity include:

• gender discordance
• separate placentas
• the so-called twin-peak sign (also called the lambda sign) (FIGURE 1), in which the placenta appears to extend a short distance between the gestational sacs; compare this with FIGURE 2, showing monochorionic twins with the absence of an intervening placenta
• an intertwin membrane thicker than 1.5 mm to 2.0 mm.

US examination can accurately identify chorionicity at 10 to 14 weeks’ gestation, with overall sensitivity that is reported to be as high as 100%.^3-5

FIGURE 1

The twin-peak sign on US

This dichorionic–diamnionic twin gestation demonstrates the so-called twin peak, or lambda, sign (arrow), in which the placenta appears to extend a short distance between the gestational sacs.

FIGURE 2

Monochorionic–diamnionic twin gestation

This US scan of a monochorionic twin gestation reveals the absence of an intervening placenta

What can go wrong

Monochorionic twins

Twins who share a gestational sac are more likely than 2-sac twins to suffer spontaneous loss, congenital anomalies, growth restriction and discordancy, preterm delivery, and neurologic morbidity.

Spontaneous loss. In 1 comparative series, the risk of pregnancy loss at less than 24 weeks’ gestation was 12.2% for monochorionic twins, compared with 1.8% for dichorionic twins.⁶ Spontaneously conceived monochorionic twins may have the highest risk of loss.⁷ However, monochorionic twins occur more often in conceptions achieved by assisted reproductive technology—at a rate 3 to 10 times higher than the background rate of monochorionic twinning.⁸

KEY POINTS

Congenital anomalies. These occur 2 to 3 times as often in monochorionic twins, and have been reported in as many as 10% of such pregnancies. Reported anomalies include midline defects, cloacal abnormalities, neural tube defects, ventral wall defects, craniofacial abnormalities, conjoined twins, and acardiac twins.^9-11 In light of these risks, a detailed anatomic survey is suggested for all twins.

 

Heart defects. The incidence of congenital heart defects is 4 times greater in monochorionic twins, even in the absence of twin-to-twin transfusion syndrome (TTTS).¹² Cardiac malformations may occur secondary to abnormal lateralization during embryogenesis or result from an abnormal vascular distribution in the shared placenta.^9,10 The presence of abnormal vascular communications may also cause limb reduction defects and the rare acardiac twin.

Long-term neurologic morbidity. In one series, the incidence of cerebral palsy was 8%, compared with 1% among dichorionic twins. In twins followed to 2 years of age, rates of minor neurologic morbidity were 15% in monochorionic twins and 3% in dichorionic twins. The overall rate of neurologic disorders in monochorionic twins was 23%, regardless of fetal weight.¹³ At 4 years, long-term neurologic morbidity was particularly high in single survivors of a monochorionic pair; the incidence of cerebral palsy has been reported to be as high as 50% in single survivors, compared with 14.3% in cases in which both twins survived.¹⁴

Twin-to-twin transfusion syndrome. In this condition, abnormal vascular connections arise in the shared placenta, allowing blood to be shunted from one fetus to the other. The syndrome is unique to monochorionic gestations and occurs in 15% to 20% of cases.¹⁵ A significant percentage of neurologic morbidity is probably the result of TTTS. To evaluate for TTTS, include a detailed anatomic survey and serial US every 2 weeks beginning in the second trimester as part of the surveillance of monochorionic twin gestations. (See “TTTS: Diagnosis, staging, treatment.”)

Down syndrome and other chromosomal abnormalities

Estimating odds

Assessing the likelihood of a chromosomal abnormality (aneuploidy) in a multiple gestation is complicated by differences in twinning mechanisms (chorionicity versus zygosity) and by the increasing rate of dizygotic twinning with advancing maternal age. The risk is greater in dizygotic twin gestations than in age-matched singleton gestations. The definition of advanced maternal age (AMA) in a twin pregnancy has ranged from 31 to 33 years of age in reports in the literature.^2,16,17

The probability that a twin gestation contains a fetus with a chromosomal abnormality is directly related to zygosity. Each twin in a dizygotic gestation carries an independent risk, so the composite risk for the pregnancy is a summation of the independent risk for each fetus. For monozygotic twins, the risk is similar to the age-related risk in a singleton gestation. Presumptions about zygosity are based on chorionicity: Almost all (90%) dichorionic twins are dizygotic and all monochorionic twins are monozygotic.

What is the utility of noninvasive screening?

Multiple gestations can be screened for aneuploidy using maternal age, maternal serum markers, and nuchal translucency (NT) on US, or combinations of these assessments.

When first-trimester serum markers (free β-human chorionic gonadotropin and pregnancy-associated plasma protein A [PAPPA]) are combined with NT and maternal age, a pregnancy-specific risk can be calculated that includes the individual contribution of each fetus, thus yielding an improved detection rate. In monochorionic twins, the NTs are averaged to calculate a single risk for the entire pregnancy. In dichorionic twins, the risk for each fetus is calculated independently and then summed to establish a pregnancy-specific risk. The combined test has a reported detection rate of 84% for monochorionic twins and 70% for dichorionic twins, compared with detection rates of 85% to 87% for singletons at a 5% false-positive rate.^18,19 The integrated test (combined test plus measurement of second-trimester serum analytes) has a 93% detection rate for monochorionic twins and a 78% detection rate for dichorionic twins, compared with 95% to 96% for singletons at the same 5% false-positive rate.^18,19 Second-trimester screening has a lower detection rate in both singleton and twin gestations.

 

Prenatal diagnosis

Given the lower detection rate of aneuploidy in twin gestations and the associated increase in aneuploidy with advancing maternal age, many patients choose to undergo prenatal diagnosis rather than relying on screening. On the basis of maternal age alone, invasive prenatal diagnosis can be offered to women who will be 31 years or older at their estimated due date.

Available diagnostic options include chorionic villus sampling (CVS) or amniocentesis. CVS is performed at an earlier gestational age (10 to 13 weeks) than amniocentesis (15 to 20 weeks). Multiples pose specific technical considerations for either procedure, and accurate fetal mapping is essential. Successful sampling with CVS can be performed in more than 99% of cases; the rate of cross-contamination is less than 1%.

Is there a risk of miscarriage?

In counseling patients about the risk of fetal death that CVS or amniocentesis may entail, the place to begin is the background loss rate, which is greater in twin than in single gestations. The reported background loss rate of twins at 24 weeks’ gestation or less ranges from 5.8% to 7.2%.^20,21 In women of advanced maternal age (35 years and older), a background rate as high as 17.6% has been described.²¹ Once parents are aware of this, they have a context for weighing the risk of miscarriage that prenatal testing may hold.

 

The twin loss rate following amniocentesis has been evaluated in several studies. (See “Integrating evidence and experience: Does invasive prenatal testing raise the risk of miscarriage in a twin gestation?”)

A greatly elevated risk of preterm birth

Multiple gestations are at extremely high risk for premature delivery, and—like all premature newborns—these infants are at risk for a wide range of disabilities. Risk factors for premature delivery include history of second trimester pregnancy loss, preterm birth at less than 35 weeks’ gestation, more than 2 previous curettage procedures, cone biopsy, müllerian anomaly, and diethylstilbestrol exposure. Unfortunately, current yardsticks for predicting premature delivery in multiple pregnancy have serious limitations, and available interventions have not been particularly successful.

Predictors

Measurement of cervical length has been evaluated as a predictor of preterm delivery in a number of twin studies that were looking for a cutoff point that can predict which twins are at greatest risk. No such cutoff has been found.^22-25

In general, studies demonstrate a low risk of preterm delivery for women who have a cervical length measurement of more than 35 mm at 24 to 26 weeks. A shorter cervical length correlates with premature delivery, but specific cutoffs have proved not to be sensitive predictors.

In the largest published series, To and colleagues evaluated 1,163 sets of twins undergoing routine care with cervical length assessments at 22 to 24 weeks’ gestation. They demonstrated a direct correlation between cervical length and preterm delivery, but were unable to define a cutoff sufficiently sensitive to be useful.²⁵ A shortened cervix may be predictive of prematurity in general, but it does not allow the obstetrician to predict with certainty which mothers will give birth prematurely or how long a particular mother will carry.

Fetal fibronectin. The presence of fetal fibronectin (ffN) in cervicovaginal secretions is widely used as an adjunct to other potential predictors of preterm delivery. In a multistudy review that included symptomatic women, a negative ffN had a 99% negative predictive value but a poor positive predictive value (13% to 30%) for delivery within 7 to 10 days.²⁶

Use of ffN in conjunction with cervical length has also been investigated in twin gestations. Although the negative predictive value of ffN remained high, the addition of ffN to cervical length assessment did not improve the positive predictive value of cervical length alone.^27,28

Interventions

Cerclage is often used in high-risk singleton pregnancies in which a shortened cervix is seen on a sonogram. The utility of cerclage in twins is less clear. Randomized controlled trials comparing women at risk of premature delivery treated with cerclage and controls not considered at risk found no difference in the rate of premature delivery in the 2 groups.^29,30 Meta-analysis of 4 randomized controlled trials also found no benefit and, in fact, detected a possibility of actual harm. Cerclage twins were more likely to deliver early (at less than 35 weeks’ gestation) and had a 2.6 relative risk of perinatal mortality. The differences found in the meta-analysis were not statistically significant, however, and the overall sample size was small (n=48).³¹

The best available data seem to show that cerclage based on US indications of cervical shortening is not beneficial and may even be associated with worse outcome.

17-Hydroxyprogesterone caproate (17P) has been found to decrease the rate of recurrent preterm birth in singleton gestations by almost 35%.³² Although twins are at increased risk of preterm birth, the use of 17P has not, however, been shown to be of benefit.³³

Bed rest. A Cochrane Database review of 6 randomized controlled trials compared 1) patients with a multiple gestation who were offered bed rest in the hospital with 2) patients hospitalized for complications of pregnancy. The review found that bed rest did not reduce the risk of preterm birth or of perinatal mortality in the routinely hospitalized women. There was, however, a tendency to a decreased number of low-birth-weight infants born to women given bed rest.³⁴

INTEGRATING EVIDENCE AND EXPERIENCE

 

Some parents elect fetal reduction

Given the high level of risk in a multiple pregnancy, reducing the number of fetuses is an option that some patients choose. A recent trend toward reduction to singleton pregnancy seems to be related to:

• increasing maternal age
• single parenthood
• financial considerations
• the increased medical risk to mother and fetuses associated with twins.³⁵

Evans and colleagues found that, although the overall rate of reduction from twins to a singleton was 3%, the percentage (76%) of women older than 35 years who opted for such a reduction was disproportionately high.³⁶

In a series of 1,000 cases of multifetal pregnancy reduction, Stone et al found that the pregnancy loss rate was lowest (2.5%) when there was reduction to a singleton gestation.³⁷ A comparative analysis of 2,000 cases of multifetal pregnancy reduction presented at the 2006 meeting of the Society for Maternal–Fetal Medicine found that the percentage of twin gestations undergoing reduction to a singleton has increased from 4% to 15.6% between 1999 and 2006, with an increase in the overall incidence of reduction to a singleton from 11.8% to 31.8%.³⁸

You should discuss pregnancy reduction with patients at high risk of pregnancy-associated complications such as cervical incompetence, preterm delivery, severe maternal cardiac disease, hypertension, diabetes, and uterine anomalies, as well as with patients who are carrying higher-order multiple gestations, in which the fetuses are at risk of problems.

Wrap-up: The tasks facing you in a multiple gestation

Start your management of a multiple gestation by taking the essential first step of determining the chorionicity of the fetuses.

Once you have done that, explain the risks of twin pregnancy and the particular risks of a single-sac pregnancy. Parents will want to know their risk of having a child with an anomaly; pay particular attention to the likelihood of a Down syndrome child. Noninvasive screening for Down syndrome and other chromosomal anomalies may be sufficient, but an older mother may prefer more definitive answers from CVS or amniocentesis.

You must prepare parents of a twin gestation for the risk of premature delivery. Cervical length assessment in the second trimester to early-third trimester may provide some indications of what is to happen, but the predictive value of this procedure is limited, and a shortened cervical length should be interpreted with caution.

In a monochorionic pregnancy, fetal growth should be assessed at regular intervals to evaluate for possible growth restriction or TTTS. If you detect evidence of abnormal fetal growth or amniotic fluid, US surveillance is indicated. Routine antepartum US surveillance of twins is not, however, recommended.²

The authors report no financial relationships relevant to this article.