Home-infusion therapy is a rapidly growing area of medicine; >3 million patients in the United States receive home and specialty infusion services anually. 1 These patients and their caregivers commonly maintain central lines in the home and perform the majority of central-line care. They are at risk of central-line–associated bloodstream infections (CLABSIs), just as patients maintaining central lines in other settings would be. Reference Keller, Williams and Gavgani2,Reference Nailon and Rupp3 Despite this risk of infection, CLABSI surveillance in home-infusion therapy is not systematically performed. Reference Keller, Alexander and Williams4
Those who investigate rates of CLABSI in home-infusion therapy, or who perform CLABSI surveillance, face many barriers, including insufficiencies in dedicated or trained staff, lack of benchmarks, and absence of structured reporting processes. Reference Oladapo-Shittu, Hannum and Salinas5 Other major barriers include a lack of clinical information and the difficulty communicating with care teams. Reference Keller, Alexander and Williams4 In hospitals, access to data (eg, laboratory results, documented signs and symptoms, and imaging and other reports) is essential, and accessible tools for performing CLABSI surveillance are needed. However, these data may be more difficult to access in home-infusion therapy settings where patients may have been admitted to hospitals with which the home-infusion staff are not affiliated.
This lack of full access to data may contribute to incomplete CLABSI surveillance and incomplete recognition of patient safety risks. Incomplete CLABSI surveillance data can affect how CLABSI surveillance data are interpreted, which has implications for CLABSI benchmarking. In other settings, such as primary care, information hazards described as contributing to patient safety risks have included the following: information underload, necessary information is lacking; information overload, too many data to organize or act on; information scatter, information in multiple places; information conflict, inability to determine which data are correct; and erroneous information. Reference Beasley, Wetterneck and Temte6 These information hazards lead to information chaos or information problems that may lead to errors. Reference Beasley, Wetterneck and Temte6 Home healthcare professionals may struggle with accessing hospital data and communicating with outpatient providers. Reference Arbaje, Hughes and Werner7 Similar challenges likely apply to home-infusion CLABSI surveillance. Characterizing information chaos in home-infusion therapy is necessary to describe the role of missing data in CLABSI surveillance Reference Malinsky, Shpitser and Tchetgen Tchetgen8 and to demonstrate the needed level of support for those doing CLABSI surveillance. 9
An understanding of contributors to information chaos could affect how CLABSI surveillance data are interpreted, including the possibility of missing data, with implications for CLABSI benchmarking in the home-infusion setting. The objectives of this study were (1) to characterize information hazards in home-infusion CLABSI surveillance and (2) to identify possible strategies to mitigate information hazards.
Materials and methods
Research approach
We conducted semistructured interviews with staff involved in CLABSI surveillance (eg, quality staff, administration, infection preventionists, etc) at 5 home-infusion agencies across the United States participating in a CLABSI prevention collaborative, including 2 in the mid-Atlantic, 1 in the Northeast, and 2 in the Midwest, covering portions of 13 states and Washington, DC. All 5 agencies are nonprofit organizations, are affiliated with academic institutions, and serve both community and academic hospitals as well as outpatient providers. Recruitment of study participants was conducted using a combination of purposive and snowball sampling, seeking respondents integral to CLABSI surveillance, and recruiting participants through the collaborative. Reference Miles, Huberman and Saldana10 We interviewed at least 2 staff members involved in CLABSI surveillance from each of the 5 agencies. Portions of the methods have been previously reported. Reference Oladapo-Shittu, Hannum and Salinas5,Reference Hannum, Oladapo-Shittu and Salinas11 This study was approved by the Johns Hopkins Institutional Review Board.
Interview procedures
Using the SEIPS 2.0 framework as a guide, we constructed a semistructured interview guide to explore systems and processes for conducting CLABSI surveillance. We included items querying the following topics: (1) work system of CLABSI surveillance; Reference Holden, Carayon and Gurses12 (2) tasks and steps for surveillance; (3) barriers and strategies to performing CLABSI surveillance; (4) experience in CLABSI surveillance; and (5) definition of CLABSI used. To ensure consistency in interviewing methods, interviews were conducted by a single team member (S.C.K.). One group interview with 4 respondents was conducted based on an interviewee request and used the same interview guide. After obtaining consent, all interviews were conducted via videoconferencing and lasted 20–60 minutes. Interviews were audio-recorded and were transcribed by a transcriptionist.
Data analyses
Transcripts were deidentified and uploaded into MAXQDA for data management and analyses (VERBI Software, Berlin, GA). The initial codebook was developed deductively from the interview guide. The deductive codes were applied to the first 3 interview transcripts. In addition, we inductively identified emergent subcodes for each parent code, resulting in a total of 38 subcodes that were applied to understand information management. Reference Beasley, Wetterneck and Temte6 This initial coding process was conducted by S.H. and S.C.K.; disagreements in coding were discussed and consensus was reached by mutual agreement. Reference Saldana13 One coder (S.H.) then coded the remainder of the interview data and both reviewers (S.H. and S.C.K.) discussed emergent findings and modifications to the coding framework. We engaged a process of constant comparison of emergent findings throughout the analytic process, and when no new codes could be identified, we considered thematic saturation to have been achieved. Reference Saunders, Sim and Kingstone14
Results
We interviewed 21 individuals between November 2020 and April 2021. The respondents were primarily nurses (N = 18, 86%) (Table 1). We describe barriers impeding the flow of information necessary for effective CLABSI surveillance (Table 2), and we discuss communication pathways required for successful CLABSI surveillance in home infusion (Fig. 1). Finally, we present recommendations for improving communication and information access in home-infusion CLABSI surveillance (Table 3).
Table 1. Participant Descriptions
a 3 did not report.
Table 2. Information Barriers and Suggested Strategies for Performing Home-Infusion CLABSI Surveillance
Note. EHR, electronic health record; CLABSI, central-line–associated bloodstream infection.
Fig. 1. Communication and information exchange pathways in home-infusion CLABSI surveillance.
Table 3. Key Recommendations for Addressing Information Chaos in Home-Infusion CLABSI Surveillance
Note. EHR, electronic health record.
Information underload
Information underload (ie, when necessary information is lacking) was common. Which patients surveillance staff needed to monitor (ie, which patients had central lines) was unclear, and it was difficult to access data required for CLABSI adjudication. Acquiring these data took time, and there was uncertainty about their quality. Information underload also resulted from difficulties communicating with hospital staff, who had details useful in determining whether a patient had been admitted with a CLABSI. Some hospital staff members did not speak to home-infusion surveillance staff because they did not understand the role of home-infusion surveillance staff in patient care.
Information underload also occurred at an agency-wide level. Although home-infusion surveillance staff typically reported CLABSIs to leadership, these data often were not shared with frontline clinicians. As a result, many frontline clinicians did not know when home-infusion CLABSIs occurred.
To address information underload, home-infusion CLABSI surveillance staff developed processes to obtain information, such as communicating with individuals at hospitals. Some used workarounds, such as using a patient or family member to get information. Others received information from triage staff or frontline clinicians making reports if a patient had a complication. Home-infusion surveillance staff needed better reports to know which patients to follow for a potential CLABSI. An infusion pharmacist leader at agency 1 explained:
“[The infusion staff] also gets reports from [the hospital system] for patients that are readmitted or have diagnosis of CLABSI and she does a reconciliation between those reports and our census … Not the most ideal workflow or process or the easiest one to manage at the moment, so we’re constantly looking for better ways to pull that information and gather that data more automatically.”
A key takeaway was to engage information technology and data analysts in developing accurate reports of duration of central lines for denominator calculations as well as accurate reports of patients with potential CLABSIs.
Information overload
Information overload (when there are too many data to organize or act on) also occurred in home-infusion CLABSI surveillance activities. At times, too much information was available, including information that was not relevant. Reports used by surveillance staff to investigate patients for possible CLABSIs were sometimes overly broad (including all patients who were hospitalized, for example) and included patients who did not have bacteremia. Many informants said that there was too much information to process when finding data, partially due to a lack of a process for tracking data and no summarization of narrative data. A pharmacist from agency 5 explained: “All of our, quote, tracking in the pharmacy system right now is just progress notes and our narrative on the patient chart.” A major outcome of such information overload was significant wasted time, in a setting where many participants felt that they did not have sufficient allotted time to complete the main tasks of their job. Some even described relevant data being overlooked.
To prevent information overload, respondents described developing clear, streamlined processes for acquiring and sharing data among members of the home-infusion agency and the home-infusion surveillance staff, incorporating shared dashboards, files, and reporting mechanisms, and evaluating critically what data are most relevant. As one nurse at agency 3 described,
“[Our EHR has] pop-ups … and… they’ve been very cluttered and information has been in there that’s not relevant to most people and, anyways, it’s caused opportunities for things to get missed, so we’[re] work[ing] on what actually should be in there.”
Information scatter
Information scatter (information in multiple places) was described as a barrier to effective communication throughout all phases of CLABSI surveillance. Home-infusion surveillance staff reported having to access information from e-mails, multiple EHRs, and telephone calls. Some staff members did not have access to EHRs of all hospitals to which their patients were admitted and had to depend on others to get information. As an IP at agency 1 stated, data in home-infusion was less standardized than hospital data: “I mean, you know how hard it is to get those reports from a standardized group of nurses that are caring for all of your patients in the hospital in one place.”
To address information scatter, an infusion pharmacist at agency 5 suggested having 1 data collection template to collect the data: “I think [a data collection tool] sounds amazing, to have 1 place to maybe just make a couple clicks here and there and be able to collect accurate data and the most complete data.” Respondents also recommended using the EHR for communication and to create reports of patients with potential complications. Interviewees suggested reducing information scatter through technologies such as videoconferencing, team-based technology, using delivery tracking tickets, shared dashboards, files, and reporting mechanisms. Another suggestion was to enable home-infusion surveillance staff access to EHRs of hospitals to which patients are commonly admitted.
Information conflict
Information conflict (ie, the inability to determine which data are correct) was another noted barrier in home-infusion surveillance. A major example of information conflict was when a clinician recorded a CLABSI in a medical record without it necessarily meeting CLABSI surveillance criteria. This commonly occurred, as clinical definitions may differ from surveillance definitions. However, many home-infusion surveillance staff took clinicians’ clinical judgement of a CLABSI for surveillance purposes. Many explained that the lack of a standardized CLABSI definition worsened information conflict, as a surveillance staff member from agency 4 explained:
“We have tried different tools that are out there to make sure that when you do CLABSIs, it has to be apples to apples, and it is really challenging … so honestly, we look at the notes by infectious disease, and infectious disease will rule out and do their thing and if they say the line is the source of the infection then we call it a CLABSI.”
Participants described the utility of the potential for a single, validated home-infusion CLABSI surveillance definition. A quality manager at agency 1 explained:
“Because we don’t have a solid definition we cannot judge whether the CLABSI is confirmed or not…. [if] that was the appropriate CLABSI definition … or not …”
Erroneous information
Finally, erroneous information (when information is incorrect) occurred in home-infusion CLABSI surveillance, leading to incorrect CLABSI reports. For example, if records were transmitted prior to cultures being complete, the surveillance staff may believe erroneously that a patient had negative blood cultures, as a pharmacist at agency 5 stated: “[Sometimes] we don’t get the final culture results, so we’re not able to put it into the actual … EMR.” Developing relationships between home-infusion surveillance staff and those with knowledge of the inpatient course of patients was one way to support surveillance staff in accessing additional patient data.
Discussion
Information chaos may contribute to decreased situation awareness about the possibility of CLABSIs. Reference Beasley, Wetterneck and Temte6,Reference Endsley15 Because agencies may be unaware of CLABSI rates, they may not act consistently to prevent CLABSIs. In addition, the increase in mental workload may lead to mistakes or burnout. Information chaos may impede accurate surveillance and requires a system to unify data collection. Mitigating this information chaos is vital to the performance of surveillance staff, the performance of the agency, and the agency’s ability to enhance patient safety by being able to address CLABSIs. We developed 5 key recommendations that could increase the accuracy and simplicity of CLABSI surveillance (Table 3).
First, we showed that home-infusion CLABSI surveillance should engage both information technology and data analyst professionals to identify which patients need to be followed for CLABSIs and which may already have a CLABSI. A study of 10 home-infusion agencies collaborating to report CLABSIs noted that several agencies did not have a pre-existing system to track complications, and few used automated reports. Reference Nailon and Rupp16 Although electronic data alone have not yet proven adequate for CLABSI surveillance, Reference Hota, Lin and Doherty17 developing reports of eligible patients is an essential part of CLABSI surveillance. Hospital EHRs contain modules to aid in identifying which patients have central venous catheters, to calculate denominator data, and to determine who has positive blood cultures. Reference Burke, Eichmuller and Messam18 Information technology (IT) and data-analyst support from hospitals and home-infusion agencies could also aid in home-infusion CLABSI surveillance. For example, IT reports used to support hospital-based infection prevention could be modified to aid in home-infusion CLABSI surveillance. However, EHRs may not always be designed to support the home health provider workflow. Reference Arbaje, Hughes and Werner7 Therefore, home-infusion CLABSI surveillance reports and tools should be developed or modified to fit the home-care workflow. Reference Gurses, Xiao and Hu19 Health system or hospital-based infection prevention departments could aid affiliated home-infusion agencies in developing these reports and tools to increase efficiencies.
Second, we identified that home-infusion agencies would benefit from shared communication structures and streamlined processes for acquiring and sharing data among team members. This would aid the home-infusion surveillance staff with getting information on patients with possible CLABSIs as well as providing information about CLABSIs to home-infusion staff. Although the EHR could be considered a central method to mitigate information chaos, Reference Beasley, Wetterneck and Temte6,Reference Arbaje, Hughes and Werner7 understanding how to design the EHR to facilitate this data exchange is essential. Reference Or, Valdez and Casper20 Team-based electronic dashboards and communication platforms may be required outside the EHR to allow team members to easily exchange IP concerns.
Third, home-infusion surveillance staff should be given access to the EHR of hospitals to which patients are commonly admitted. EHR data are essential to understanding the hospital course of a patient diagnosed with a complication, and these data are also important in home-health information management. Reference Arbaje, Hughes and Werner7 However, home-infusion patients can be admitted to hospitals with different EHR systems. Some EHR platforms and even states allow data sharing through health information exchanges. 21,22 Home-infusion agencies should be provided access to these health information exchanges.
Fourth, we identified the importance of a validated, benchmarked home-infusion CLABSI surveillance definition. A 2008 home healthcare–associated bloodstream infection definition 23 relied on National Healthcare Safety Network (NHSN) acute-care CLABSI surveillance criteria in use at the time, 24 but it has not been widely adopted. Reference Rinke, Bundy and Milstone25 A proposed home-infusion CLABSI surveillance definition still requires validation prior to larger use. Reference Keller, Salinas and Williams26 A validated, widely accepted, home-infusion CLABSI surveillance definition will enable communication among home-infusion surveillance staff.
Fifth, we learned that home-infusion CLABSI surveillance staff could benefit from building relationships with individuals at hospitals, such as healthcare epidemiologists or infection preventionists who may be able to provide information or context. Others have noted the need to bring hospital-based and home-based staff together to understand information needs and to redesign information-sharing tools between hospital-based and home-based staff. Reference Arbaje, Hughes and Werner7,Reference Nasarwanji, Werner and Carl27,Reference Romagnoli, Handler, Ligons and Hochheiser28 We suggest that this information redesign should also occur in home infusion therapy.
Our findings have implications for calculations of CLABSIs rates in home infusion therapy. Each of the information barriers identified could lead to missing data, which could lead to either an underestimation or overestimation of the CLABSI rate. This misestimation may affect perceptions of CLABSIs in home infusion therapy. In a national survey, some home-infusion professionals stated that they did not see CLABSIs as a problem in their setting. Reference Keller, Alexander and Williams4 Our study is one of the first analyses of information management in HAI surveillance. Others could use these models to investigate information flows in HAI surveillance in other settings.
Our study had several limitations. The participants were all members of 5 nonprofit home-infusion agencies in the mid-Atlantic, Midwest, and Northeastern United States, each of which is affiliated with or owned by an academic health system, which increases the likelihood of selection bias and limits generalizability. Specific tasks required for CLABSI surveillance may be different in for-profit agencies, agencies in other parts of the country, or in those not affiliated with an academic health system. However, participants worked in home-infusion agencies in 13 different states. Because we focused primarily on CLABSI surveillance, we did not focus interviews on those not directly involved in surveillance, so we interviewed fewer agency leaders or frontline clinicians. In addition, because these agencies have agreed to participate in a collaborative focused on CLABSI surveillance and prevention, participants may have been more experienced with CLABSI surveillance than other agencies. It is possible that information management practices may be different at nonparticipating agencies.
We suggest a focus on mitigating inadequate communication structures by increasing access to information technology or data analyst specialists, improving intra-agency communication structures, increasing access to hospital EHRs. Also, a validated home-infusion CLABSI surveillance definition and closer relationships with home-infusion surveillance staff and inpatient personnel are needed. We suggest that the possibility of missing data be considered in analyzing the burden of home-infusion CLABSIs. Finally, we suggest a greater focus on home-infusion CLABSI surveillance to understand the rates of CLABSI in home infusion therapy. Health systems should work to support the information flows required for appropriate home-infusion CLABSI surveillance to ensure that CLABSI prevention activities can be initiated.
Acknowledgments
We appreciate the dedication of the Home Infusion Collaborative members in participating in this initiative.
Financial support
This research was supported by the Agency for Healthcare Research and Quality (grant no. R01HS027819 to S.C.K.).
Conflicts of interest
All authors report no conflicts of interest related to this article.
Open access
