A MODEL FOR PREVENTING ACCIDENTALFALLS IN HOSPITALS
-MANAGEMENT PLAN FOR BRIEF CHANGE IN PATIENT CONDITIONS-
Shogo KATO1, Satoko TSURU2, Yoshinori IIZUKA3
123Department of Chemical System Engineering, The University of Tokyo, Japan, JSQC,
Hongo 7-3-1, Bunkyo-Ku, Tokyo 113-8656, Japan
1The University of Tokyo,
2The University of Tokyo,
3The University of Tokyo,
ABSTRACT
These days, more interest is focused on quality and safety of healthcare. In particular, prevention and reduction of medical accidents is essential. Medical accidents are roughly classified into two types. Onetype is accidents caused by healthcare staff in the process of providing healthcare services, and the other is accidents caused by patients who receive medical services (“accidents caused by patient state”).
Among accidents caused by patient’s state, accidental falls remain a significant problem. While “assessment score sheets” have already been used in hospitals to prevent and reduce accidental falls, satisfactory results have not actually been achieved.
In this study, we aim to establish a methodology for preventing accidental falls based on a strategy of “identifying high-risk patients and their high-risk situations, and carefully allocating healthcare staff to manage patients’ behavior to prevent the identified dangerous situations.”
The ‘management plan’ for each patient that realizes this strategy includes three factors. First, a plan of instructions for patients on actions they can take for safety in their daily living is essential as a base (“basic plan”). Second, a plan to keep up with any short term change in a patient’s state is needed, because the state of a hospitalized patient will usually be temporarily affected by medication and changing disease conditions (“corresponding plan”). Third, an additional plan to deal with noncompliant patients is important, because patients sometimes do not follow the instructions of the medical staff and take undesirably dangerous actions (“additional plan”).
In this paper, we develop a model for preventing accidental falls, which enable us to design appropriate management plan for each patient. Then, we develop a prototype system based on the designed model. Finally, we address the result of verification of the model, by applying the prototype system into actual cases in hospitals.
Keywords: medical accident, healthcare safety, quality assurance
1. INTRODUCTION
1.1 Background
These days, more interest is focused on quality and safety of healthcare. The prevention or reduction of medical accidents is particularly essential. Medical accidents are roughly classified into two types. One type is accidents caused by medical staff in the process of providing healthcare services, such as medication errors and accidents due to misunderstanding the patient. The other is accidents caused by patients who receive medical services (“accidents caused by patient state”), such as accidental falls, evulsionsinvolving tubes or continuous infusion needle, patients leaving the hospital prior to discharge, and accidental self-mutilation.
Among accidents caused by the patient’s state, accidental falls remain a significant problem. Such accidents account for 16 percent of total medical accidents and are the second largest accident category[1](Ministry of Health, Labor and Welfare, 2001). Sometimes they also have important consequences for a patient [2] (Fujimoto, 2003), such as fractures.
Assessment sheets[1], [3] (Sugiyama et al., 2005), aiming to prevent and reduce accidental falls, have been suggested as tools for assessing risk of patient falls, and they are used in hospitals. An example of an assessment sheet, which has been suggested by the Japanese Ministry of Health, Labor and Welfare[1], is shown in Figure 1. On these sheets, scores are distributed for each risk factor questionnaire item, and we can assess the fall risk of a patient by adding the scores of the checked items indicated on the sheets.
However, satisfactory results have not actually been achieved. While there are some known reasons for not achieving satisfactory effects, we consider it impossible to specifically identify dangerous situations with these sheets. Questions such as when is it dangerous, what actions are dangerous for the patient, and other related questions are not addressed with these tools, because the output of the sheets is the overall risk for the patient. Therefore, it is difficult to adopt appropriate countermeasures to prevent accidental falls based on the informational output of these sheets.
Figure 1-Example of Assessment Sheet
1.2 Strategy of this study and previous studies
Accidental falls, as with any other kind of accident caused by patient state, occur mainly when patients take undesirable actions, those actions the medical staff have not assumed. Having sufficient medical staff to guide and control patient movement is the most efficient approach for preventing accidental falls. The next need is to appropriately allocate human resources, because human resources in hospitals are always limited. Finally, it is important to keep up with any short term changes in patient states, because the state of a hospitalized patient will usually be temporarily affectedby medications and changing disease conditions (these are“influencing factors”). Therefore, in this study, we adopt the following strategy for preventing accidental falls. First, we identify both the high-risk patients and their high-risk situations, taking into consideration possible short term changes in patient state. Next, we carefully allocate medical staff for the identified dangerous situations.
To identify the high-risk patients and their high-risk situations, we can refer to “a process model for determining elderly care for the Activity of Daily Living (ADL) (shown in Figure 2),” proposed by Kato et al. [4] (Kato et al., 2008). The specificknowledge contents of the model have already been developed and the adequacy has already been confirmed [5] (Kato et al., 2007). Using this model, we can determine a pattern of realization for each ADL item as a combination of element actions (specific actions broken into component steps to achieve a certain ADL item) and specific care.
Based on these considerations, Kato et al. [6] (Kato et al. 2006) proposed a model for preventing accidental falls. Through retrospective verification, using actual cases, the adequacy of the model for cases without influencing factors was confirmed. However, also clarified was that the model would not work for cases with influencing factors. We need to consider further both the behavior of influencing factors and the overall design of the model.
1.3 Purpose of this study
Considering the background and strategy in Section 1.1 and Section 1.2, the goal of this study is to develop a hospital sickbed management system that enables hospitals to appropriately allocate medical staff for the prevention of accidental falls.
In this study, we aim to develop a model for preventing accidental falls that enables us to design an appropriate management plan for each patient, which can become the basis of a sickbed management system.
To achieve this goal, in this paper we first identify the fundamental functions and concepts needed to prevent accidental falls, based on our strategy (Chapter2). Then, we design a model for preventing accidental falls, based on these previously identified functions and concepts (Chapter 3). Finally, we develop a prototype system with the required databases (or tools), based on the designed model (Chapter 4).
2. FUNDAMENTAL FUNCTIONS AND CONCEPTS NEEDED TO PREVENT ACCIDENTALFALLS
In this chapter, we identify the fundamental functions and concepts needed to prevent accidental falls.
First, we describe the fundamental concept of Plan, Do, Check, Act (PDCA) cycle management based on management plans for each patient on the ward (Section 2.1). Then, we identify the factors needed to design a management plan for each patient through an analysis of actual cases in hospitals (Section 2.2). Lastly, we describe the fundamental functions and concepts used to design the factors we identified (Section 2.3).
2.1 PDCA cycle based on management plan for each patient
Operations for preventing accidental falls are usually conducted by nurses on the ward, based on a management plan for each patient as one of day-to-day work. Nurses manage each patient by keeping up a PDCA cycle from the inception of the management plan for each patient (shown in Figure 2). The management plan needs to include not only the care to be provided to the patient by the medical staff but also instructions for the patient, noteworthy points on the patient’s state during surgery, and other details.
2.2 Three factors needed for each patient management plan
To identify the factors needed to design a management plan for each patient, we analyzeda total of 23 actual cases occurring between April 2006 and September 2006 in two hospitals (7 cases in X hospital and 16 cases in Y hospital), and discussed the factors with the chief nursesin charge of preventing accidental falls in each hospital.
As a result, we believe we can enumerate three situations in which accidental falls occur in hospitals;
Situation 1: Accidental falls occur when patients take dangerous actionswhile attempting to achieve certain daily living purposes.
Situation 2: Accidental falls occur when a patient’s mental or physical state has changed in the short term,due to the effects of medications or changing diseaseconditions, and in attempting to achieve certain daily living purposes the patient takes actions that are dangerous in the patient’s current state.
Situation 3: Accidental falls occur when the patient takes unexpected anddangerous actions against medical staff instructions.
Based on the above, we believe that the management plan for each patient needs to include three factors. First, a plan of instructions for patients on actions they can take for safety in their daily living is essential as a base (“basic plan”). Second, aplan to keep up with any short term change in a patient’s state is needed, because the state of a hospitalized patient will usuallybe temporarily affected by medications and changing disease conditions (“corresponding plan”). Third, an additional plan to deal with noncompliant patients is important, because patients sometimes donot follow the instructions of the medical staff and take undesirably dangerous actions (“additional plan”).
Figure 2-PDCA Cycle based on Management Plans for each Patient
2.3 Fundamental functions and concepts needed to prevent accidental falls
The model for preventing accidental falls we aim to design hastwo objectives. We identified the fundamental functions needed to achieve each purpose, as shown in Table1, by consulting previous studies [4] [6]. The details of each function and concepts needed to achieve each function are described in the following section.
Table 1-Functions for Preventing AccidentalFalls
2.3.1 Designing thebasic plan
• Function1-1: Assuming actions of a patient
Accidental falls are considered to occur when patients take certain actions. We therefore consider that controlling such actions taken by a patient is essential to preventing accidental falls (??intended meaning okay? ??). Behaviors of hospitalized patients are generally limited, and according to accident descriptions, ADL accounts for a large number of incidence for accidental falls. Therefore, in this study, all patient actions are assumed to be ADL.
Generally, there are many ways to realize each ADL item. We need to break each ADL item down into specific actions to express the variety of ways each ADL item can be realized. In this study, we term these specific actions “element actions.” Element actions are specific actions that have been broken into component steps to achieve a specific ADL item. That way we can express a variety of ways to achieve specific ADL items as multiple “patterns of realization,” which is a combination ofelement actions to achieve a specific ADL item.
• Function 1-2: Assessing fall risk for assumed actions
After assuming the actions a patient may take, the fall risk involved in the assumed actions is assessed. In this study, we assess the fall risk by evaluating it at the following three levels.
• safe: a level where the patient can achieve the action by him/herself with low-risk
• some risk: a level where the patient can achieve the action by him/herself yet still incurring some risk
• need some assistance: a level where the patient can achieve the action with low-risk,due to receivingassistancefrom someone
To evaluate the three levels, we introduce ability elements as evaluation indicators fora patient’s state. The patient’s actual condition is expressed as a score for each ability element. We call the indicators“actual abilities.” In addition, each element action determines “required abilities,” which represent abilities necessary to achieve the element action in each of the three levels. Eventually, we can evaluate the fall risk for each element action by comparing actual abilities with required abilities.
• Function 1-3: Determining basic instruction
Based mainly on the output of function 1-2, the safe pattern of realization to be actually achieved is specified for each patient. Also specified is whether or not medical staff assistanceis needed. In addition, the bed rest level, which expresses the area the patient is allowed to move based on medical judgment, is usually indicated by the charge doctor. Eventually, a basic plan for the patient is indicatedwith these three items.
2.3.2 Designing the corresponding plan
• Function 2-1: Assessing influencing factors
A hospitalized patient’smental and physical state will usually betemporarily affected by medications and changing disease conditions (these are“influencing factors”). How the patient’s state will change depends on the types of influencing factors and the patient’s state before hospitalization. Also, how the medical staff should handle the patient depends on the type of influencing factors and the comprehensive ability of the patient. We need to assess these points for each patient.
• Function 2-2: Determining additional instruction and staff assistance
Based mainly on the output of function 2-1, additional instructionsaregiven to the patient. Additional instructions consist of a discussion of both the risksposed by influencing factors and safe patterns of realization, considering the influencing factors. In addition, special monitoring rules consisting of check points and appropriate reassessments are also needed to avoid medical staff overlooking influencing factors and changes in the patient’s state.
If the patient’s state is expected to be temporarily lowered by the effects of influencing factors, we need to instruct the patient in safer patterns of realization or instruct the patient to take action only with medical staff assistance.
If the patient’s state is expected to recover from the effects of the influencing factors, we particularly need to instruct the patient to take actions only with medical staff assistance, because the patient’s ability to move is reducedduring the stay at the hospitaland we need to be careful due to the change in the patient’s state. Eventually, a corresponding plan for the patient is indicated by these items.
2.3.3 Designing the additional plan
• Function 3-1: Assessing detailed feature of a patient
For noncompliant patients, who donot follow medical staff instructions and take undesirably dangerous actions, particular patienthandling is needed to prevent accidental falls. To determine additional plans, we assess the detailed features of the patient: comprehensive ability, degree of overconfidence, actual abilities, frequency with which they takeundesirable dangerous actions, what undesirable actions the patients tend to take, and other related information.
• Function 3-2: Determining additional correspondence for a patient
Based mainly on the output of the function 3-1, we determine how to handle the noncompliant patient. First, we consider preventing accidental falls by using special resources such as low beds or special bed railings, or special tools that control patient’s action such as control bands. If it is difficult to completely prevent accidental falls, we need to consider countermeasures for early detection or effect mitigation. Eventually, an additional plan for the patient is indicatedwith these items.
As above, a management plan for each patient is designed.
2.3.4 Designing management plan for the ward
• Function 4-1: Estimating required resources for each patient
Based on the management plan for each patient, the required resourcesfor each patient are estimated. Required resources for a patient areboth the amount of human and physical resources necessaryfor the patient management plan.
•Function 4-2: Considering available resources on the ward
Required resources on the ward arethe number of human and physical resources essential for the ward to administer all management plans for patients by totaling all resources required for each patient. In addition, available resourceson the ward are the available number of human resources for ADL assistance and physical resources on the ward. Eventually, we can assess the feasibility of the management plan for each patient by comparing required resources with available resources on the ward.
Though the value of available resources on the ward should be treated as an approximation, if the available resources on the ward are considerably less than the required resources, we need to review the management plan for each patient.
As above, the management plan for the ward is designed.
3. PROPOSING A MODEL FOR PREVENTING ACCIDENTALFALLS
3.1 Method for designing a model for preventing accidental falls
For a model thatactualizesthe concepts and functions described in Section 2.3, we designed a model for preventing accidental falls that consists of the procedures and databases (or tools) needed to achieve the objectiveby following four steps.
(1) We identified the input and output of each function described in Section 2.3.
(2) We identified what items we should prepare before performing each function, without depending on the individual abilities of the users.
(3) We identified the databases (or tools) needed during the performance of the function, based on the items we identified in (2).
(4) We designed detailed procedures and structure for the databases (or tools), based on conventional output.
3.2 The model for preventing accidental falls
The entire picture of the model for preventing accidental falls is shown in Figure 3, and the procedures and the databases (or tools) in the model are shown as Table 3.