Context-aware framework and
4.4 SMART WARD SYSTEMIMPLEMENTATION
document which is considered as service advertisement. By using the same domain ontologies by both the service requestors and providers simplifies the matching process.
D. Service Registry
Unlike the traditional web service registries, the semantic web service registry contains references to semantic information annotating the advertised services.
E. Matching Algorithm
These algorithms are designed so as to help in matching the semantics of provided service descriptions and requests and are in general more complex and intelligent that the syntax-based method.
The main difference between the traditional architecture and the semantic web service architecture is the implementation and the integration of the semantic matching engine with the service registry.
75
Second, data is fed by the patient care professionals after measuring the readings from medical devices which are not either capable of transmitting it electronically or expensive devices which cannot be dedicated for each bed.
For instance, a treatment ward can have only one Arterial Blood Gas (ABG) analyser machine. In such case, the system has to depend on the entries manually made.
There are many research work results which point to the techniques of data cleansing, pruning, aggregation, and interpretation. Many standards have evolved for converting and presenting raw sensor and actuator data to machine readable and interpretable form. For this research work, such data is simulated using controlled random values. Currently, the system is designed to act on the contextual changes triggered by the assumed sensor based values.
Figure 4.4 PHP code for Drain service implementation
Figure 4.4 is the code written for a drain service. Nusoap is used to create and implement the client services and server services. Figure 4.5 is part the code for the server. The WSDL file which is used in service composition is also developed using Nusoap. Figure 4.6 Illustrates the WSDL file needed for discovering the service and use in the composition.
Figure 4.5 PHP code for server implementation
Figure 4.6 Extracts from the WSDL file
Context Attributes are built from the literature available on the various measurements and its threshold / reference range as recommended by medical practitioners and medical literatures available (Caroline and Céline, 2010; Royal Prince Alfred Hospital Patient Observation (Vital Signs) Policy Adult (2010); Pat (2001). These attributes are specific for the post cardiac surgery. A complete list of attributes for the healtcare domain would be exaustive and requires professionl assistance. Rules for determing goal sitiation is defined based on these attributes. For this implementation the context space is limited to an ITU.
77
Table 4.2 Context Attributes Context
Attributes Measured in Reference Range as
Quality value (Quantity value) Internal
Temperature
○C
low (< 36)
mid (>= 36 and <= 37) high (> 37)
Surface Temperature
○C
low (< 35)
mid (>= 36.5 and <= 35) high (> 36.5)
Transcutaneous oxygen
saturation [SpO2]
%
poor (< 80)
fair (>= 80 and < 90) good (>= 90 and < 98) excellent (>= 98) Arterial
Blood Pressure mm Hg
low (< 90/70)
mid (<= 130/90 and >= 90/70) high (> 130/90)
Drain level ml < 150ml
Respiratory Rate breaths/minute
normal (12-20) low (<8) high (>24)
Level of
Consciousness AVPU
Alert. The patient is alert and interactive Voice. The patient responds to voice.
Pain. The patient responds only to central pain.
Unresponsive. The patient is unresponsive.
Fibrinogen mg/dL
high(>400) normal 200–400) low(<200)
Table 4.3 Sample Data sheet
Internal_temperature in C Surface temperature in C Transcutaneous oxygen saturation [SpO2] in % Arterial Blood Pressure mm Hg Systolic Diastolic Pulse beats/min Fibrinogen globulin Drain_Qty in ml Respiratory Rate breaths/minute Level of Consciousness Blood Glucose Level
36 35.6 99.3 mid 127 90 100 390 38 276 19 P 130
36.3 38.8 98.3 high 180 96 100 470 31 170 28 A 155
39.4 34.5 98.6 high 146 96 93 429 33 293 25 A 87
36.1 34.8 96.7 high 136 96 93 466 31 230 26 P 146
37.6 37.9 99 high 178 95 93 327 37 262 24 U 105
39.4 34.3 97.8 mid 118 87 93 128 36 203 18 U 89
37.7 35 99.5 high 158 91 96 463 38 208 28 P 145
39.6 37.7 99.5 mid 127 75 93 272 37 251 12 P 80
37.5 38.1 96.9 mid 117 87 92 289 39 205 16 V 144
37.8 34.2 99.4 mid 92 87 92 482 41 123 13 P 90
37.1 36.6 90.1 mid 94 73 92 253 36 233 16 A 95
37.3 34.1 97.5 high 178 95 92 196 29 123 25 A 120 38.6 39.4 99.3 high 150 95 93 254 36 288 25 U 156
37.7 37.5 96.6 mid 97 88 90 147 34 139 17 A 97
36.2 35 92.6 high 151 98 90 356 38 237 26 P 120
38.5 37.2 98.6 high 167 95 89 489 37 106 26 V 86
34.9 37.3 92 high 137 92 89 365 30 148 28 U 179
36.1 34.3 95.5 high 175 93 89 169 46 165 27 U 135
37.9 38.9 95.2 mid 110 79 89 490 36 158 12 V 104
38 36.9 96.5 mid 129 70 90 274 39 291 19 U 156
35.5 38.5 97.1 mid 116 75 88 486 31 280 14 V 143
34.4 37.9 98.9 high 141 96 90 310 31 179 26 A 150
37.6 37.2 92.1 mid 114 74 90 220 28 278 16 U 126
37.2 37.4 96.3 mid 109 80 90 293 30 123 15 A 79
39.9 38.5 90.4 mid 116 74 89 111 36 147 18 P 112
34.2 36.1 90.8 mid 103 80 89 483 35 226 16 U 97
39.4 36 94.9 low 81 67 88 432 37 156 12 A 150
34.2 38 99 high 137 94 87 290 32 137 24 P 162
37.4 37.6 98.5 mid 98 83 87 123 37 278 18 P 130
37.3 37.1 92.8 mid 100 87 87 449 46 221 18 V 132
36.7 37.2 93.3 high 134 95 87 159 44 116 27 V 173
34.8 35 92 mid 108 89 87 334 32 133 16 P 151
79
Table 4.2 lists the context attributes identified while Table 4.3 is a compilation of sample data representing various scenarios that could emanate in an ITU. The implemenation was tested with five hundred records to check for dynamic service composition. Only thirteen parameters that are very critical for the use case proposed were considered Caroline, Arbour, and Gélinas Céline, (2010). These parameter are for a cardiac surgery.
However, the same parameters might be present for other healtcare systems.
Table 4.4 Johns Hopkins Cardiac Surgery and post-surgery team Cardiac Surgeons
Cardiac Resident Physicians Cardiac Anesthesiologists Cardiac Anesthetists Cardiac Intensivists
Cardiac Physician Assistants Cardiac Surgical Assistants
Cardiac Nurse Practitioners Cardiac Perfusionists
Mechanical Circulatory Support (MCS) Coordinators
Cardiac Transplant Coordinators Cardiologists
Critical Care Fellows Operating Room Nurses Operating Room Technicians Intensive Care Unit Nurses Progressive Care Unit Nurses
Typically a number of doctors and other personals, apart from numerous machines and equipments, are involved in patient care at various stages. For a cardiac surgery, the people involved, the procedures to be followed and a
complete information of patient care from pre-operative preparation to post- operative care is given by the Johns Hopkins Medicine hospital. Table 4.4 is a summary of the patient care personals including the operation team and the other assistance involved in a cardiac surgery as listed in the webpage.
(http://www.hopkinsmedicine.org, 2015)
4.5 ALGORITHM FOR SERVICE COMPOSITION TO ACHIEVE