Improving Post-operative Pain Management Using the CIS Model

• Abstract
• Introduction
• Background to Project
• Process
• Discussion
• Conclusion
• References

Abstract
Currently it is difficult to quantify the direct benefits that Clinically Integrated Systems (CIS) have on patient care delivery. However, when new systems and processes are developed and underpinned by a computerised programme to support their implementation, the interrelationships and complexity of care delivery can be uncovered. The following case study outlines how a CIS can have a positive impact on patient and organisational outcomes.

Introduction
Historically, utilisation of information technology within health care organisations has focused on administrative or billing tasks as opposed to clinical activities.^[1-3] Unfortunately, the perceived benefits of implementing these systems to help contain spiralling health costs have not been achieved.^[4,5] It is now recognised that the most significant health costs are initiated at the point of clinical services delivery but a lack of computerised clinical information systems (CCIS) has meant that this information is not being captured.^[6,7]

Paradoxically, while clinicians want information systems to help improve health care delivery, they are also actively resisting its introduction.^[8,9] This dichotomy is easier to understand if the purposes of such information systems are clarified. Clinicians are not averse to working with sophisticated information technologies that aid in the diagnosis or treatment of patients, for example, radiography equipment. Familiarity with and use of computers is now part of daily clinical practice - even if it is limited to accessing personal email. The problem arises when CCIS are proposed.

A growing body of literature highlights a high implementation failure rate in health care organisations that have attempted to introduce CCIS.^[10-12] The reasons for this include the complexity of these systems^[13] and a lack of organisational change management programmes to support these new initiatives.^[14,15]

In 2000, the Clinically Integrated System (CIS) Model of CCIS was developed by the A+ Network Centre for Best Patient Outcomes ("the Centre"). The CIS Model is an interdisciplinary electronic system that links evidence based care, clinical redesign, outcome management and participatory action research into a single framework to improve patient care (figure 1).

Figure 1: The patient is at the centre of the CIS Model. Encasing the patient is an "A" framework with a gold star that represents clinicians’ commitment to excellence in care delivery.

At the tips of star are the philosophies of care that are interlinked. The tips of the star facing outward is a reminder that the CIS Model must continually evolve to meet the needs of patients and clinicians.

The CIS Model has been designed to represent the care provided for any patient group regardless of condition or diagnosis. As the Model collects data, information about both individuals and patient groups can be analysed.

The following case study report outlines the patient, clinical, organisational and financial benefits of a CCIS using the CIS Model, specifically in this example, how it was used to to address adverse clinical variance in "poor postoperative pain management". The ramifications of this project are then discussed.

Background to Project
In 2000, the CIS Model prototype was introduced in the Orthopaedic Department at Auckland City Hospital (ACH) for patients with fractured neck of femur (#NOF). In 2004, a comparison study was undertaken. All patients admitted to ACH with #NOF from 1 January 2000 to 30 June 2004 were divided into two groups: Group A consisted of patients who received their health care via the CIS Model and Group B consisted of patients who did not. The results showed a reduced median length of hospital stay (LOS) of 3 days for the CIS Model patients (Group A) as opposed to Group B (figure 2). Analysis showed that patient age, sex or surgeon were not confounders in the difference in median LOS.

Figure 2: Comparison of average LOS for Group A and Group B

To help explain this difference, a more detailed analysis of the data for Group A patients was undertaken. That analysis examined the impact of the clinical changes that had been implemented at the point of patient care delivery for Group A patients. A significant finding related to pain management.

From 2000 to 2001, the CIS Model outcome trends indicated that 32% of #NOF patients had received poor post-operative pain management. In 2002, a clinical project was instigated to address this problem. By 2004, the incidence of this variance had reduced by 27% (figure 3).

Figure 3: Post-operative pain trends for #NOF patients

Process
The Centre’s Orthopaedic Interdisciplinary Team (IDT) initially presumed that the high pain variance incidence was associated with poor IDT assessments and inadequate medication regimens. However, the CIS Model demonstrated that the IDT’s assumption was incorrect.

In-depth analysis of the data showed that while nurses were offering regular analgesia to patients, a high percentage of patients were declining medications based on a variety of misperceptions regarding pain relief.

This finding was supported by data uncovered through a literature search undertaken by the IDT, which showed that many elderly people decline medications for reasons such as fear of addiction, undesirable side effects, fear of needles and a reluctance to report pain to staff.^[16-18]

The IDT took a multifaceted approach to solving the problem. The first step was to develop an Evidence Based Pain Guideline (EBPG) with a strong emphasis on patient education. Using the CIS Model Evidence Based Guideline (EBG) generic template, a comprehensive EBPG was developed that included information on pain definitions, evaluation of pain assessment scales, and recommended interventions.

The Centre’s previous experience had shown that while an EBG template is a beneficial teaching aid and reference document, the length of most guidelines needs to be adapted to ensure that information is pertinent to individual patient needs. Consequently, from the reference EBPG, short treatment intervention statements were modified and collated so that clinicians could select appropriate treatments and directly download this information into the individual patient record (figure 4).

Figure 4: An example of evidence-based pain statements that have been modified to allow the IDT to select interventions and directly download information on interventions into the individual patient record

Once the intervention information has been downloaded, a number of options become available to the clinician, eg, "continuing" the intervention, which requires no additional entry, or, alternatively, entering an "update" progress note. When treatment is discontinued or is completed, an outcome(s) must be selected prior to leaving the screen (figure 5).

Figure 5: An example of an EBG statement, update note on the patient progress and final outcomes

Recommendations drawn from the EBPG have been adapted into a one-page information sheet for patients. This sheet can be printed out on request (figure 6).

Figure 6: A sample of the patient information fact sheet on pain relief

In conjunction with the EBPG, a formal process for evaluating patient pain levels was introduced. Nurses are to complete a pain assessment for each patient at regular intervals. Trends in assessment "scores" over time assist clinicians to decide when a satisfactory level of pain control has been reached (figure 7).

Figure 7: Example of a pain assessment scoring trend

As the pain data are collected, the information is automatically collated for analysis from patient population perspective. This allows data analysis on the types of EBG interventions being used and their overall effectiveness in patient care delivery (figure 8).

Figure 8: Example of monthly evidence-based statements used for pain education from a patient group perspective; the type of treatment intervention is linked to the number of patients who used it and the effectiveness outcome recorded.

The impact of the pain variance on patient care delivery is not viewed in isolation. The pain data were integrated into the CIS Model monthly reports (figure 9), which can link demographics, clinical variances, financial data, and patient and relative/whanau participation outcomes together.

Figure 9: An example of an automated monthly outcomes report for #NOF patients

When the pain variance data were analysed in conjunction with other outcomes a relationship between a patient declining pain medication and a subsequent failure to meet the expected mobilisation criteria emerged. In turn, this finding was linked to an increased LOS (table 1). The effectiveness of this project became apparent from January 2002, in a measurable decline in average LOS for #NOF patients with pain.

Table 1: Improving yearly trend of the average length of stay and additional costs for #NOF patients with unresolved pain problems compared to those without pain

*Extra cost = Extra days stay x No. of patients x $580

Discussion
The impact of computerisation on patient care delivery can be clearly demonstrated by the introduction of the CIS Model into clinical practice. While it is beyond the scope of this paper to discuss the organisational and clinical challenges associated with the development of the CIS Model prototype, the project management strategy used during this phase helped to lay the future foundations for clinicians’ acceptance to not only continue to use the Model, but also to act on unfavourable trends in outcomes.

In this project, the CIS Model demonstrated that the IDT’s assumption that the pain variance was related to poor practice by clinicians was incorrect by providing data that highlighted patient compliance as the major problem. The ramifications of this finding are wider than the initial project scope, because they reveal a trend that the traditional form of documentation with pen and paper has failed to uncover. The CIS Model program has a component that concurrently captures patient feedback in a qualitative data form. As the feedback in this free text accumulates, it can be quantified. For example, the discovery that patients’ declining pain relief was a large factor in pain variance led to the addition of a new outcome field to the program to support the continual tracking and quantification of the overall incidence of patients declining treatment. Consequently, an immediate feedback loop between the use of recommended evidence-based guidelines, treatment interventions, outcomes and the patient’s perspective had been created for all variances.

The dissemination and implementation of new clinical information and practice can be problematic, particularly where a large and diverse group of health professionals is involved. In this project, two additional features were added to the CIS Model program to resolve these issues. First, all members of the IDT were individually informed of the introduction of the new pain guideline using a computer messaging system when they logged onto the CIS Model.

Secondly, a software feature was designed to ensure compliance with the guideline. While the majority of clinicians completed the necessary documentation, a small number of "offenders" had been identified through clinical pain spot audits - by matching entry omissions and the individual staff member assigned to care for the patient. Consequently, a completion reminder message system was developed. An automated reminder message sent to the individual staff member when they logged on to the computer has helped resolve the problem. This messaging feature is now being extended to include a teaching component that will provide individual feedback on well-written documentation or where areas of improvement are required.

A reassuring trend has been the move towards a decline in pain variances with a coinciding rise in patient education (figure 10).

Figure 10: Comparison of incidence of pain variance with pain education variance

The positive changes emerging during the project have resulted in a recent request by clinicians for the CIS Model to be used to provide guidelines on newer pain relief techniques, such as epidurals in postoperative hip and knee replacements. This is seen as a significant advance by the Centre members as it represents a change in focus. When the CIS Model was introduced into clinical practice, the emphasis of debate initially centred on the issues of computerisation. However, with the production of useful clinical data, the focus is now moving towards improving patient outcomes, which was the purpose of introducing computers.

To help further encourage the dissemination of the latest evidence-based guidelines, the CIS Model program has been extended to include self-learning modules, like the pain guideline, which on completion can provide a clinician with "credits" towards yearly performance reviews.

With the introduction of the CIS Model prototype for #NOF patients, the Model has been implemented for single diagnostic related conditions, such as elective hip and knee surgery patients and has been extended to service-based implementation, for example, Adult and Paediatric Palliative Care Services. The #NOF EBPG has subsequently been modified by those working in these areas to meet the specialised needs of their patients. This adaptation of an existing approach for use in other areas has proved to be cost-effective and supports the cross-fertilisation and sharing of information across patient populations.

Conclusion
The project demonstrates how the integration of new clinical practices supported by computerisation can have a direct, beneficial impact on patient health care delivery. The CIS Model was able to uncover previously unknown links between pain and its wider effects on the financial, organisational, clinical and patient aspects of health care delivery.

References

1. Snyder-Halpern R, Chervany, N. A clinical information system: strategic planning for integrated healthcare delivery networks. J Nurs Adm 2000;30(12):583-591.
2. Bates DW. The quality case for information technology in healthcare. Medical Informatics and Decision Making 2003;2(7):1-20.
3. Rowe I, Brimacombe P. Integrated care information technology. NZ Med J 2003;116;1169:1-8.
4. Mckee M, Aiken L, Rafferty A, Sochalski J. Organisational change and quality of health care: an evolving international agenda. Qual Health Care 1998;7:37-41.
5. Littlejohns P, Wyatt JC, Garvican L. Evaluating computerised health information systems: hard lessons still to be learnt. Br Med J 2003;362: 860-863.
6. Anderson JD Increasing acceptance of clinical information systems. MD Comput 1999;16(1):1-8.
7. Panko WB. Clinical care and the factory floor. J Am Med Inform Assoc 1999;6:349-353.
8. Barr BJ. Managing the change during an information systems transition. ACORN 2002;76(6):1085-1092.
9. Dienemann J, Van de Castle B. The impact of healthcare informatics on the Organisation. J Nurs Adm 2003;33(11):557-562.
10. Southon FCG, Sauer C, Dampney CNG. Information technology in complex health services: organisational impediments to successful technology transfer and diffusion. J Am Med Inform Assoc 1997;4:112-124.
11. Staggers N, Thomas CR Happ B. An operational model for patient - centered informatics. Comput Nurs 1999; 17(6):278-285.
12. Van der Meijden MJ, Tange HJ, Troost J, Hasman A. Determinants of success of inpatient clinical information systems: a literature review. J Am Med Inform Assoc 2003;10:235-245.
13. Doonlan DF, Bates DW, James BW. The use of computers for clinical care: a case series of advanced US sites. J Am MedInform Assoc 2003;10:94-107.
14. Kirley D. Not whether, but when: gaining buy in for computerised clinical processes. J Nurs Adm 2004;34(2):55-58.
15. Kirley D, Stein M. Nurses and clinical technology: Sources of resistance and strategies for acceptance. Nurs Econ 2004; 22(4):216-224.
16. Carr E. Refusing analgesics: using continuous improvement to improve pain management on a surgical ward. J Clin Nurs 2002;11(6):743- 752.
17. Victor K. Properly assessing pain in the elderly. RN 2001;64(5):45-49.
18. Chodosh J, Ferrell BA, Shekelle PG, Wenger NS. Quality indicators for pain management in vulnerable elders. Ann Intern Med 2001;135(8):731-735.