Software requirements: definition and specification.

The software requirements specification is the single most important document in the software development process. It provides the basis for development as well as for validation. The SRS needs to include adequate definition of all requirements without specifying implementation or project management issues. The SRS should be completed early in the development process. However, it is very likely that changes will occur during the development life cycle. This is not an excuse for approving and releasing the current version of the SRS. When changes occur, the SRS must be revised. In any case, the concept is to deal with the current, approved version of the SRS. Ultimately, the SRS should include all the information needed to proceed into the design phase of software development.

Temporal patterns of radiographic infiltration in severely traumatized patients with and without adult respiratory distress syndrome.

We prospectively evaluated the patterns of pulmonary structural and functional changes in 100 consecutive surgical intensive care unit trauma patients who had (1) emergent major surgery, (2) a pelvic fracture, or (3) two or more major long bone fractures. For each patient, arterial blood gas measurements (ABGs), central venous pressure (CVP), pulmonary capillary occlusion pressure (PAOP), thoracic compliance, arterial oxygen tension/fraction of inspired oxygen (PAO2/FIO2), pulmonary venous admixture (Qs/Qt), and portable chest roentgenograms were sequentially tracked. The senior staff radiologist interpreted all chest roentgenograms. Pulmonary infiltration was quantitated in each of six fields using a scale ranging from 0 to 4, with 0 being no infiltration and 4 being the maximum. Adult respiratory distress syndrome (ARDS) was defined as follows: Qs/Qt > or = 20%, PAO2/FIO2 < 250 or both; dependence on mechanical ventilation for life support for > or = 24 hours; PAOP or CVP or both < 20 mm Hg; and thoracic compliance < 50 mL/cm H2O. Time zero (T0) the time of onset of ARDS, was defined as the time these criteria were met. Eighty-three of 100 study group patients had penetrating injuries, and 17 were admitted with blunt trauma. Fifty-one of 100 patients developed ARDS: 36 of 51 died. Only 4 of 49 (8%) patients without ARDS died. The injured lungs of patients with and without ARDS had similar amounts of infiltration over most measured time intervals. The noninjured lungs of the ARDS patients, however, had significantly greater infiltration than those without ARDS at T0 and over subsequent time intervals.(ABSTRACT TRUNCATED AT 250 WORDS)

The relationship between ARDS, pulmonary infiltration, fluid balance, and hemodynamics in critically ill surgical patients.

Hypervolemia from fluid overload with resultant pulmonary edema is thought to be a frequent cause of Adult Respiratory Distress Syndrome (ARDS). However, ARDS may also occur as a result of the hypovolemic shock of surgery or trauma. To develop an appropriate rationale for fluid therapy in high-risk surgical patients, the relationship between fluid balance, hemodynamics, the onset of ARDS by physiologic criteria (shunt greater than or equal to 20%, and/or PaO2/FiO2 ratio less than 250) and the onset of pulmonary infiltration (PI) associated with ARDS were examined. Fifty patients were prospectively followed from admission throughout their hospitalizations; 38 (76%) had trauma and 12 (24%) were postoperative. Cardiac index, central venous pressure (CVP), wedge pressure (WP), and shunt (Qsp) were measured. All chest x rays were read by one staff radiologist who was blinded to the patients' identities. PI was graded from "0" to "4" (0 = no PI, 4 = maximum PI). The first x ray reading of "2" or greater was used as the time of onset of PI. ARDS by physiologic criteria occurred in 29 of 50 (58%) patients; 27 of these 29 (94%) also developed +2 or greater PI. The mean onset times of ARDS and of +2 PI were 40 +/- 41 hours and 40 +/- 38 hours, respectively. The ARDS patients had a significantly smaller net positive fluid balance than the non-ARDS patients over the first 40 hours after admission (+6,831 ml +/- 4,909 ml vs 12,440 ml +/- 7,817 ml, (P less than 0.01)).(ABSTRACT TRUNCATED AT 250 WORDS)

Software quality assurance and system safety.

The medical field is faced with the increasing sophistication of software-controlled medical equipment. Clinical engineers must understand and become familiar with these sophisticated software-driven medical devices and stand-alone medical software programs in order to evaluate their suitability for the desired purpose. They must be able to assess the nature of the software, its potential and its risks. Numerous aspects of software development and safety must be considered when developing and/or evaluating such software devices and programs.

Experimental renal artery thromboembolism.

The natural history of experimental renal artery embolism with autogenous thrombus in 16 mongrel dogs has been investigated by sequential selective renal angiography and correlated with examination of the kidneys following autopsy. Angiography in the early (0-24 hr) post-thromboembolism period shows evidence of arterial obstruction and/or spasm which progressively disappears leaving no apparent permanent damage. However, within one or two weeks arterial and nephrographic changes and loss of renal size are discernible by careful comparison with the pre-embolism angiogram. At autopsy multiple disseminated microscopic infarcts involving the convoluted tubules were demonstrable in all cases. Occasionally large wedge-shaped infarcts involving the full thickness of the cortex were superadded. The potential importance of these findings relating to angiographic technique and assessment of treatment of renal thromboembolism is discussed.

The effect of glucagon on barium-enema examination.

Fifty barium-enema studies were performed with glucagon and 50 with a placebo to compare their effect on colonic spasm, patient discomfort, and diagnostic quality. Each drug was administered in a randomized double-blind fashion and was injected intramuscularly 10 minutes before beginning the enema. Bowel relaxation during fluoroscopy was graded. Patients were questioned about discomfort during and immediately after the enema, and radiographs were reviewed blindly for diagnostic quality and degree of spasm. Studies done with glucagon produced significantly less spasm and discomfort and better diagnostic quality compared to the placebo (p less than 0.01).