Objective. To evaluate an Internet-based telemedicine program designed to reduce the costs of care, to provide enhanced medical, informational, and emotional support to families of very low birth weight (VLBW) infants during and after their neonatal intensive care unit (NICU) stay.
Background. Baby CareLink is a multifaceted telemedicine program that incorporates videoconferencing and World Wide Web (WWW) technologies to enhance interactions between families, staff, and community providers. The videoconferencing module allows virtual visits and distance learning from a family's home during an infant's hospitalization as well as virtual house calls and remote monitoring after discharge. Baby CareLink's WWW site contains information on issues that confront these families. In addition, its security architecture allows efficient and confidential sharing of patient-based data and communications among authorized hospital and community users.
Design/Methods. A randomized trial of Baby CareLink was conducted in a cohort of VLBW infants born between November 1997 and April 1999. Eligible infants were randomized within 10 days of birth. Families of intervention group infants were given access to the Baby CareLink telemedicine application. A multimedia computer with WWW browser and videoconferencing equipment was installed in their home within 3 weeks of birth. The control group received care as usually practiced in this NICU. Quality of care was assessed using a standardized family satisfaction survey administered after discharge. In addition, the effect of Baby CareLink on hospital length of stay as well as family visitation and interactions with infant and staff were measured.
Results. Of the 176 VLBW infants admitted during the study period, 30 control and 26 study patients were enrolled. The groups were similar in patient and family characteristics as well as rates of inpatient morbidity. The CareLink group reported higher overall quality of care. Families in the CareLink group reported significantly fewer problems with the overall quality of care received by their family (mean problem score: 3% vs 13%). In addition, CareLink families also reported greater satisfaction with the unit's physical environment and visitation policies (mean problem score: 13% vs 50%). The frequency of family visits, telephone calls to the NICU, and holding of the infant did not differ between groups. The duration of hospitalization until ultimate discharge home was similar in the 2 groups (68.5 ± 28.3 vs 70.6 ± 35.6 days). Among infants born weighing <1000 g (n = 31) there was a tendency toward shorter lengths of stay (77.4 ± 26.2 vs 93.1 ± 35.6 days). All infants in the CareLink group were discharged directly to home whereas 6/30 (20%) of control infants were transferred to community hospitals before ultimate discharge home.
Conclusions. CareLink significantly improves family satisfaction with inpatient VLBW care and definitively lowers costs associated with hospital to hospital transfer. Our data suggest the use of telemedicine and the Internet support the educational and emotional needs of families facilitating earlier discharge to home of VLBW infants. We believe that further extension of the Baby CareLink model to the postdischarge period will significantly improve the coordination and efficiency of care.
- neonatal intensive care unit
- very low birth weight (infant)
- randomized controlled trial
- medical informatics
Broadband: A high-speed Internet connection
ISDN: Integrated Services Digital Network provides high-speed digital phone service that can be delivered through standard phone lines and facilitates videoconferencing
Mb, Gb: Megabyte and gigabyte, respectively. A measure of the amount of memory contained in a computer
RAM: Random access memory
Shockwave Applet: A computer application used within a World Wide Web (WWW) browser to display animated and interactive content
Virtual house calls: The use of videoconferencing technology to allow clinicians to check on a patient without the need for a home visit
WWW browser: A program such as Microsoft Internet Explorer or Netscape Navigator that allows a computer user to view information contained on the World Wide Web
Secure Sockets Layer: A standard method for encrypting data during transit across the Internet
The admission of a newborn child to the neonatal intensive care unit (NICU) is among the most emotionally distressing situations that a family can face. Young families with little experience with critical care medicine are thrown into a high-tech environment that is bewildering and foreign to most parents.1 A sick child may pose emotional, educational, and logistic problems for a parent. During a family's stay in the NICU and after discharge, they need not only top-quality medical care, but also effective and creative information sharing. We hypothesized that Internet and telemedicine technologies could influence these family needs and lower health care costs. In this report we will describe our efforts to design, implement, and test a high-tech approach to provide individualized support to the families of very low birth weight (VLBW) infants.
Telemedicine broadly defined means care at a distance. Early telemedicine approaches could almost be characterized as medicine with 2-way television.2 The convergence of voice, data, and video on the broadband Internet lends support to the notion that telemedicine can and should use this new medium for transmitting and obtaining clinically relevant information. Almost 120 million Americans have accessed the Internet from home and office. The use of the World Wide Web (WWW) to gather medical information has also skyrocketed with over 45 million Americans using the WWW to gather health information.3 Fast access to the Internet within the home is happening more rapidly then any other technological introduction in modern history. Despite the ubiquitous availability of the Internet in the home, the use of the Internet to evolve new models of care is unexplored. Here we report the first randomized, controlled clinical trial of this technology to change the delivery of care.
Baby CareLink was created by a team of neonatologists, nurse practitioners, nurses and respiratory therapists, social workers, child life specialists, medical informatitions, and software engineers. Baby CareLink provides information to families using both a specially designed WWW-based system and a system of videoconferencing from the NICU. The CareLink system is programmed using Microsoft BackOffice (Microsoft Corporation, Redmond, WA) components including Internet Information Server 4.0, Active Server Pages and SQL Server 6.5. Security Services were provided using ACE Server (RSA Security Incorporated, Bedford, MA). Baby CareLink dynamically generates WWW pages that can be accessed from a standard web browser. Educational content is enhanced with video, audio, and ShockWave applets (Macromedia Company, San Francisco, CA).
Six major areas of clinical content and resources are present within Baby CareLink Web including a daily clinical report, a message center, a see your infant section, a family room, a clinical information section, and a section focused on preparation for discharge to home (Fig 1). The daily report is a web-page that provides clinical updates about an infant's clinical care and status. This report relies on dynamic links to the NICU's electronic medical record system (HP CareVue, Hewlett Packard Company, Andover, MA) to avoid the need for clinicians to enter information into 2 separate systems. The message center is a WWW-based messaging system through which parents can share confidential communications with members of the NICU staff. Baby CareLink also contains a context sensitive messaging throughout the CareLink site to allow parents to easily compose messages related to the content they are viewing. The see your infant section is a pictorial daily journal comprised of images captured by the staff with a consumer grade digital camera. Baby CareLink also provides a mechanism for allowing families to share these photographs outside of the confines of the CareLink security architecture. By changing a picture's status through the WWW-based interface, parents can post pictures to a password protected WWW where their families and friends can see their infant. This approach maintains the locus of control for making these photographs public with the parents. The family room provides a potpourri of supports including answers to common questions, information about services available to families, links to WWW-based resources, an on-line library for browsing available print and video resources. “The Kid's Corner” provides a collection of information and support materials specifically geared for older siblings of our patients. “The Emotional Side of the NICU” allows parents to both read about the issues that confront families of high-risk newborns and view high-quality digital video of NICU families discussing how they coped with their NICU experience. The clinical information and care section describes the issues present at various stages in an infant's NICU stay including when an infant is first admitted, as an infant stabilizes and family members becomes more active participants in their infant's care, and the period before discharge when families prepare to take over all of their infant's care at home. The “NICU-Pedia” provides an on-line encyclopedia of clinical conditions, tests, treatments, and medications relevant to the care of high-risk newborns. The preparing for discharge section is an on-line discharge teaching module where parents can view multimedia modules describing the knowledge and competencies they must acquire before discharge. This module is constructed so that NICU clinicians can individualize the content for each infant's needs using a simple WWW interface. It also allows parents and clinicians to track acquisition of knowledge.
Patients enrolled in the evaluation study intervention group received a standard Hewlett Packard Vectra 233 MHz Pentium II processor computer with 32 MB of RAM, a 3.1 GB hard disk, a 12X CD-ROM drive, a 33.6 baud full duplex fax modem with a speakerphone, a 16-bit sound card, and a 17-inch SVGA color monitor with integrated speaker. The homestation, which has Windows 95, was equipped with PictureTel (Andover, MA) Live 200 desktop videoconferencing equipment. Families could access the WWW-based information from any computer equipped with a standard web browser. Connections to CareLink were made through standard phone lines for WWW access, while videoconferencing used 128Kb ISDN lines. Both the computer equipment and communication lines were provided to families at no cost to them. Arrangements for installation of home units were made at the time of randomization. Units were left in the families' home for 4 to 6 months after discharge. An outside company installed the equipment in each family's home and retrieved it at the end of the study period. Clinicians in the NICU had Internet access through hospital clinical workstations. Three mobile videoconferencing carts with a PictureTel group conferencing system allowed conferencing from each of the NICU's 33 bedspaces.
The identity of users is verified using RSA SecurID hardware tokens (RSA Security Inc, Bedford, MA). These tokens are small, handheld devices containing microprocessors that calculate and display unpredictable, one-time-only codes. The codes automatically change every 60 seconds. To access patient-specific data, the CareLink site requires that a user begin a session by entering a username and a password. The password is the combination of a memorized personal identification number (PIN) and the currently displayed 6-digit code on the RSA SecurID device. This information is transmitted to a security server that authenticates the user and verifies that the correct password was entered. The security server compares the user-entered password with its knowledge of the correct password for that 60-second period. If the password does not match, it also checks the password from the previous minute to account for delays in typing and transmission. Once a password is verified, the user is authenticated for the entire CareLink site for the duration of the web session or 15 minutes, whichever is less. Each user is assigned to a group (parent, NICU staff, community provider, or administrator) and is linked to 1 or more patients. This information is initialized at the beginning of a session and then used for the entire session. CareLink dynamically generates web pages based on the group identifier and patient links of the user. CareLink uses the group identifier to provide different features on the web pages and uses patient links to limit patient data access to patients linked to that user. The RSA SecurID username, password, and all patient data are transmitted through the standard Secure Sockets encryption Layer to protect confidentiality. Extensive audit logs are kept for each session.
Project participants were selected from VLBW infants born at Beth Israel Deaconess Medical Center, Boston Massachusetts, between November 1, 1997, and March 30, 1999 and cared for in its NICU. Infants were excluded if ISDN access was not available at their family's primary residence or if the infants expected length of stay in the study NICU was expected to be <14 days (eg, because of need for transfer for surgical care at the nearby Children's Hospital, or because of planned transfer back to a referring community hospital for infants born weighing >1250 g). In addition, infants were excluded if their family lacked a permanent residence, did not speak English, or if discharge to other than the biological family was expected. Lack of basic telephone service in the family residence was not used as an exclusion criterion. For such families, basic telephone service was arranged for the duration of the study period. Attending physicians could exclude a family from the study if they felt enrolling the family in a study would be clinically inappropriate. The families of eligible study infants were approached for consent to participate between the infant's third and tenth hospital day.
After informed consent was obtained, infants were randomized to the intervention or control group using a birth weight-stratified permuted block design. Infants in the control group received information and support as usually provided in the NICU, while those in the intervention group were given access to the Baby CareLink system as described above. In cases of multiple deliveries (ie, twins/triplets etc), 1 infant was randomly selected for random assignment to either the intervention or the control group. Treatment status of the remaining siblings was yoked to that of the selected sibling (ie, if the infant was assigned to the intervention arm, his or her siblings also received the intervention). Only data about the randomized infant is included in analyses. The clinical team made all decisions regarding clinical care including those regarding the timing of discharge.
Sociodemographic and clinical information was collected from the medical record and from family interviews. Recorded elements included birth weight, gestational age, sex, plurality, maternal race, mode of delivery, and Apgar scores as well as maternal characteristics including maternal educational level, gravidity, parity, and marital status. Disposition from the study NICU was recorded as 1) alive and discharged from the hospital, 2) alive and transferred to another facility (eg, level II, III or rehabilitation facility), or 3) died. Discharge date from the study hospital was recorded, as was the date of ultimate discharge home for those transferred to other facilities. Information on the frequency of family visits and telephone calls to the NICU as well as episodes of kangaroo care and other types of holding by the family was abstracted from the medical record. Information on a family's previous experience with the use of computer technologies was recorded at the time of family training using a structured written survey instrument.
Each family in the Baby CareLink group was given a single training session that focused on the hardware and software to be used in their home. These sessions lasted between 45 and 120 minutes with most lasting <75 minutes. The local phone company installed ISDN lines in the family residence and computer hardware was placed and tested by a local hardware service provider. Hardware and ISDN lines were placed in most homes within 12 days of randomization. In only 1 case, installation required more than 3 weeks.
The Picker Institute's Neonatal Intensive Care Unit Family Satisfaction survey4 was used to assess family perceptions of the quality of care provided to them and their infant by the NICU. This 80-item written questionnaire was administered to families between 1 and 4 months after their child's discharge from the NICU. Surveys were not sent to families of infants who died in the NICU or whose child was discharged to chronic care facilities. The Picker instrument is designed to assess family experiences within the following 8 dimensions of care: 1) information and education to parents, 2) environment and visitation policies, 3) family and infant support by the NICU, 4) confidence and trust in the NICU, 5) continuity and transition, 6) family participation in care, 7) overall impressions, and 8) coordination of care. Responses to each question and within each dimension are tabulated to form a problem score, representing the percentage of questions within a dimension that elicited a problem response. For example, Item 67 asks, “How would you rate the way the NICU staff worked together? Possible responses included poor, fair, good, very good, and excellent. The first two of these are characterized as problem responses.
Statistical analysis was performed using the SAS Statistical Software version 6.12 for Windows (SAS Institute, Cary, NC). Simple comparisons were done using χ2 testing or Wilcoxon rank sum tests where appropriate. Length of stay was compared with the Wilcoxon rank sum test. Additionally, proportional hazards models were fit, allowing adjustment for birth weight and the ability to treat in-hospital deaths as incomplete (ie, censored) observations. Because the more complicated model did not change the conclusions, only the Wilcoxon test results are reported.
Comparison of problem scores for both individual questions and dimensions was done using Fisher's exact test.
Descriptive analysis of web utilization was performed using 2 methods. Family and staff sessions were identified and tabulated using the CareLink Security logs. Utilization of individual sections and pages within CareLink was identified from the web server logs maintained by Internet Information Server. Because the breadth of WWW content changed during the course of the study, we analyzed a 6-month period during which all content was present and was unchanging (November 7, 1998, through April 7, 1999). This tabulation was done using Web Trends Log Analyzer Version 5.0 (WebTrends Corporation, Portland, OR). Utilization of videoconferencing was assessed by review of the ISDN line billing records.
The Beth Israel Deaconess Medical Center's Committee on Clinical Investigation approved the study.
One hundred seventy-six VLBW infants were admitted to the NICU during the study period. Eighty-eight infants (50%) were excluded from the trial because of the presence of 1 or more exclusion criteria (Table 1). Eighty-eight infants were eligible for randomization. The families of 9 infants declined to participate and 4 were unavailable to provide consent during the enrollment period. Of the 75 remaining infants, 26 were randomized to the Baby CareLink arm and 30 were randomized to receive usual care. The 7 siblings of infants in the Baby CareLink group and 12 siblings of control infants were placed in the same study group as their randomized sibling.
The 120 excluded infants were of slightly higher birth weight (1087 ± 286 vs 995 ± 290 g; P < .05) and higher gestational age (28.8 ± 2.7 vs 27.7 ± 2.3 weeks;P < .05) than the 56 study infants. The excluded infants and 56 study infants did not differ in gender distribution, plurality, maternal race, or insurance status.
Demographic and clinical characteristics of the 2 study groups are presented in Table 2. The 2 groups were similar in birth weight, gestational age, plurality, maternal race, and educational level, as well as insurance status. Infants in the Baby CareLink group were more likely to have been delivered by cesarean section (92.3% vs 63.3%; P < .05) and to be born to a single mother (38.5% vs 13.3%; P < .05). Similar percentages of parents of each group lived together in the same home (88.5% vs 90.0%; P ≥ .85).
Infants in the CareLink group remained hospitalized for a shorter time than those in the usual care group although this difference was not significant (68.5 ± 28.3 vs 70.6 ± 35.6 days;P ≥ .05). Interestingly, posthoc subgroup analysis showed the difference among infants born weighing <1000 g to be even greater (77.4 ± 26.2 vs 93.1 ± 35.6 days; P≥ .05). The difference in lengths of stay among infants of higher birth weight was smaller (54.4 ± 26.9 vs 48.2 ± 16.8 days;P ≥ .05) Of note, control infants were significantly more likely to be back-transported to level II facilities (20% vs 0%;P < .05).
During the study period, 1033 CareLink WWW sessions were initiated by NICU and project staff members (1.8 sessions/day). During the same period, the 26 CareLink families initiated 1744 sessions. This represents 67 sessions per family and an average of 0.98 sessions per inpatient day. The average duration of WWW sessions for all users was 5.4 minutes. Most family sessions were initiated from the home. The mother's security token was used to initiate 64% of the WWW sessions.Table 3 shows that the patient-specific areas within the CareLink WWW were the most commonly visited. During the study period, families initiated 328 videoconferencing sessions. These sessions lasted on average 6 minutes.
Postdischarge surveys were administered to 51 families. Five families were not surveyed because of NICU death (1 family) or transfer to and discharge from another level III or rehabilitative facility (4 families). Responses were received from 31 families (61% response rate). The response rate was similar in both groups. There was no difference between responders and nonresponders when compared on birth weight, gestational age, Apgar scores, length of stay, and maternal age or educational level.
As can be seen in Fig 2, CareLink usage was associated with significant improvements in family satisfaction in the overall quality of care and environment and visitation dimensions. CareLink families also reported higher scores in all other dimensions except in coordination of care. Within the dimension of overall quality, (Fig 3A) CareLink families were 85% less likely to report problems with the duration of their child's hospitalization (6.7% vs 43.8%; P = .04). Of those reporting problems most noted that their NICU stay was shorter than they felt necessary. Interestingly, even though the same visitation policies applied to both groups, CareLink families were also less likely to report problems when asked if the unit's visitation policy met the needs of their other family members (13.3% vs 50%;P = .02; Fig 3B). CareLink families also showed a trend toward fewer problems related to receiving practical support from the NICU (33.3% vs 68.7%; P = .08).
We have demonstrated that emerging communication technologies such as the WWW and videoconferencing can be successfully integrated into practice within a busy NICU and that they significantly improve family perceptions of care quality and lower costs associated with transfers. NICUs and the care of VLBW children are costly, constituting the single largest segment of hospitalized care for children. Previous studies have shown that improved home care lowers costs for VLBW infants.5 Yet home care by appropriately trained nurses is expensive and logistically difficult in many settings. We believe extension of Baby CareLink into the postdischarge period can provide this type of support to families and the clinical staff and will significantly improve the coordination and efficiency of care.
Families and staff alike enthusiastically embrace the use of the Baby CareLink system. Parents used the system daily mostly looking at their infant's WWW home page and the photograph gallery. Some content in Baby CareLink such as the NICU-Pedia's section on retinopathy of prematurity were less frequently accessed. However, given the small number of families in our study (n = 26) and the rarity of the condition, this finding is not surprising. Moreover, the total number of accesses to a section should not be equated to the potential impact even a single reading might have for the family. From the staff perspective, they now have adopted Baby CareLink as part of the care for each child in the NICU. Children routinely have postdischarge virtual visits, and families have experimenting with virtual support groups.
The ability and willingness of families in the NICU to use the Baby CareLink WWW to gather information about their child's care in the NICU parallels the experience of Bass et al.6 In their study of elderly patients with Alzheimer's disease they found a computer support network to be heavily used by the patient's family caretakers. In addition, they found that the system significantly reduced stress among caregivers. Although our project facilitated access to the WWW by providing computer hardware and access to the Internet, our surveys show that 70% of families by the end of our study already have home-based WWW access. These figures are similar to that found by other Boston-based research,7,8 but is higher than the 38% found nationally in 1999.
Others have also demonstrated the ability of telemedicine technologies to be valuable in the care of pediatrics patients. Karp et al9 in their study of a population of children with special health care needs found a telemedicine system to be both well-received and well-utilized in providing distance consultation within the state of Georgia. The use of telemedicine to provide post-intensive care unit (ICU) home monitoring and care has been described by Miyasaka et al.10 Using a system of videophones in the patient's home and the pediatric ICU, they were able to demonstrate reduced need for physician home visits, unscheduled hospital visits, and days hospitalized. Despite these successes, obstacles do exist to the routine implementation of systems such as these. These issues include the ability of providers to recoup costs associated with providing telemedicine consultations,11medicolegal issues,12 as well as issues related to access.8
Almost all families can learn to use CareLink, which is designed for a 6th grade reading level. The lack of access to the Internet by many families poses logistic problems for the NICU wishing to adopt this kind of approach. In our study, we had governmental support that provided equipment and telecommunications into the home. Work by Gustafson (personal communication, April 2000) has shown that patients without traditional access to health care derive greater benefit from computer-assisted decision support. Homes with televisions and telephones can be made Internet-ready for as little as $200. Learning centers at the hospital can help provide access when there is none at home. Local libraries can also provide needed access for families as well.
The small size of our study group limits the power of our analyses to detect differences. Our analyses do suggest that Baby CareLink improved the transition from NICU to home. Although their NICU stays were shorter, parents in the Baby CareLink group expressed greater comfort with the timing of their infant's discharge. We believe that Baby's CareLink allows families to attain the skills, knowledge, comfort, and confidence necessary to become the primary caretaker for their child(ren) earlier in their NICU stay. Furthermore, extension of the Baby CareLink model and technologies into the postdischarge period and enhancing communication between families, community, and tertiary providers may further improve care.
Families in the CareLink group reported greater satisfaction with care across multiple dimensions. In addition to the improvements documented by the results of our Picker surveys, we believe the decreased rates of retrotransfer among CareLink families was the result of improved satisfaction with the care received in the study NICU. In this NICU, it is a part of standard procedures to retrotransfer infants back to their referring level II facility when they no longer require level III care. This process occurs unless families actively decline this transfer.
To our surprise, families in the Baby CareLink group expressed greater satisfaction with the NICU's physical environment and visitation policies despite the fact that both groups were cared for in the same NICU and experienced the same set of clinical policies. Although not immediately self-evident, it appears that this results in part from Baby CareLink's ability to facilitate alternative methods of visiting for extended family members. For example, families often used videoconferencing to allow daily visiting of young children with their hospitalized siblings. Our experience suggests that such “real world” visiting is often limited by the need for parents to divide their attention between their hospitalized and older child.
The ability to support low birth weight infants (<1000 g) is a marvel of modern medicine, but extracts an emotional toll on the family. Our study suggests that our intervention designed to support the emotional and educational needs of families actually improved care and supported families who felt comfortable taking their children home sooner. Families without CareLink technology tended to stay longer but felt that their stay was too short. Technologies such as the ones we have evaluated can offer every family better communication with their physicians, better coordination of care with their nurses, and better collaboration with other members of their community. Paradoxically, we have used a high-tech approach in a high-tech environment to provide a more humane high-touch.
This study was funded by the National Library of Medicine's Telemedicine Initiative (NO1-LM-6-3535).
We would like to thank the members of the Baby CareLink Study Advisory Panel for their guidance during the course of this project. Panel members included Heidelise Als, PhD, Mary Ellen Avery, MD, Patricia Flatley Brennan, RN, PhD, Jerold Lucey, MD, and Marie McCormick, MD, ScD.
We would also like to thank Ms Beth Hinton for her help with manuscript preparation.
- Received May 26, 2000.
- Accepted July 3, 2000.
Reprint requests to (J.E.G.) Beth Israel Deaconess Medical Center, Department of Neonatology, 330 Brookline Ave, Boston, MA 02115. E-mail:
Disclosure: Dr Safran is Chief Executive Officer of Clinician Support Technology (CST). Ms Pompilio-Weitzner is currently clinical content specialist to CST. Dr Gray holds equity in and serves as a consultant to CST. CST is a developer and distributor of CareLink applications.
The Baby CareLink Clinical and Technical Team consists of the following individuals: Sheleagh Alsop-Somers, MSW; Hollis Caswell, RN; Peter Jones; Alison Levy, MS; Glen Low, RRT, PhD; Alfredo Morales, MS; Michele Phillips, RN, PhD; Mary Quinn, RN, NNP; Annette Roberts, RN; David Veroff; Qiang Wang, MD; and Gail Wolfsdorf, MSW.
- NICU =
- neonatal intensive care unit •
- VLBW =
- very low birth weight •
- WWW =
- World Wide Web •
- PIN =
- personal identification number •
- ICU =
- intensive care unit
- Wasserman W. Complications. New Yorker. February 28, 2000:87–109
- Field MJ, ed. Telemedicine: A Guide to Assessing Telecommunications for Health Care. Institute of Medicine; 1996
- Peterson A. Lost in the maze: the web's explosive growth poses a challenge for users: making the most of it. Wall Street Journal. December 6, 1999, Section R:6 (column 1)
- The Picker Institute. Improving the quality of health care though the eyes of patients: Surveys. 2000. Available at URL: http://www.picker.org/surveys/default.htm
- Brooten D,
- Kumar S,
- Brown LP,
- et al.
- Bass DM,
- McClendon MJ,
- Brennan PF,
- McCarthy C
- Goldsmith D, Safran C. Using the web to reduce postoperative pain following ambulatory surgery. Proc AMIA Symp. 1999;780–784
- Mandl KD, Katz SB, Kohane IS. Social equity and access to the World Wide Web and E-mail: implications for design and implementation of medical applications. Proc AMIA Symp. 1998;215–219
- Karp WB,
- Grigsby K,
- McSwiggan-Hardin M,
- et al.
- Miyasaka K, Suzuki Y, Sakai H, Kondo Y. Interactive communication in high-technology home care: videophones for pediatric ventilatory care. Pediatrics. 1997;99(1). URL: http://www.pediatrics.org/cgi/content/full/99/1/e1
- Sorum PC, Mallick R. Physicians' opinions on compensation for telephone calls. Pediatrics. 1997;99(4). URL: http://www.pediatrics.org/cgi/content/full/99/4/e3
- Copyright © 2000 American Academy of Pediatrics