This research work develops a diagnostic and feedback-providing system for self-learning by volunteers who are asked to log in to a website to study the required subjects in three weeks.

The system establishes students' learning weaknesses and offered guidelines for their improvement until all such weaknesses are corrected.

Assessment is based on students' relative growth scores, and their record profiles.

Website development: The www.nmac.in.th website contains the learning courseware database. The program is interactive, with personal files for each student along with other details like code, password, name, surname, date-month-year, test scores, and log-in periods and time.

Students can, on their own, plan their learning for each subject; they are free to record and copy the articles on the website. There is no log-in limit from any place with internet connection around the clock.

Learning courseware: We developed a set of learning courseware of the specified topics as activities and diagnostic tests for arterial blood gases (ABGs).

The activities covered topics under the table of specification, whereas the diagnostic tests probed the weaknesses of students' understanding of activities so as to assess their learning skills in analytical thinking of patients' problems under assorted circumstances to be faced in clinical practice.

Each set of courseware contained three levels: beginner, intermediate, and advanced.

Students are required to pass each in an orderly way, at least once with a minimum score of 80% before advancing to the next level.

Once they have passed certain courseware, they are free to review all learning courseware as often as they wanted to.

Learning sources: Interactive with students, the program identifies learning weaknesses while avoiding giving them direct answers immediately after each test.

What it does is instead to link the student to begin a search for the well-prepared content, which contains details of the subject.

Construction of test forms: Pre-test and post-test forms are designed to be parallel under the table of specification, thus preventing the carry-over effect by testing them three weeks afterward and alternating the questions or changing them.

The tests underwent verification for content validity and had their index of item objective congruence (IOC) determined by four clinical experts.

Deciding whether the tests and behavioral objectives aligned, they set the IOCs at 0.79 for pre-tests and 0.73 for post-tests.

• Two sets of pre-tests and post-tests of 10 questions each, with each question containing four sub-items requiring short answers
• 15 activities and 10 diagnostic tests of the multiple-choice (4) questions

System testing: We introduced the elements, namely the learning courseware, tests, and sources of learning, to the developed system before testing it under the research and development approach, the logical approach, and empirical approach.

Different sample sizes were used, and logical comments were analyzed to test the system efficiency under three stages:

First-stage system development and assessment:

• The sample group was 4 sixth year medical students who underwent system testing during a three-week period
• Objectives:
• To test system operating units and system log-ins
• To test the objectivity of the tools
• Findings:
• The system operating units were flexible to students' interactive behavior, instruction responsiveness, and network connection. Access to the system through the website provided ready access to the information, was user-friendly, and was stable.
• The tools were objective, covered complete details, and were clear, with elements of the subject content in proper order.
• Improvement:
• Making the system interactive with the network; focusing on conciseness; adding channels for answering questionnaires through the website; adding the record on downloading articles from the website; notifying students of their performance on the tests, question by question; and enabling the website to accommodate no less than 100 concurrent users.
• Adjustment of fonts to match general personal-computer sets and updating the wording of the questions to be more relevant with the circumstances of patients.

Second-stage system development and assessment:

• The sample group was 32 nurse anesthetist students and 21 first year residents in anesthesiology who underwent system testing during a three-week period.
• Objectives
• To test criterion-referenced tool quality
• To determine learning achievement
• Findings:
• The pre-test and post-test registered 0.48 and 0.50 in item difficulty (p) values; 0.72 and 0.73 in sensitivity indices (s); 0.85 and 0.88 in content validity; concurrent validity through calculation of Pearson Product moment against a standard test of 0.82; and 0.91 reliability for the internal consistency by the Kuder-Richardson method 20.
• The diagnostic tests registered 0.87, 0.91, 0.87 and 0.93, 0.93, 0.86 in item difficulty (p) values; 0.17, 0.11, 0.20 and 0.15, 0.08, 0.21 in sensitivity indices; 0.89, 0.80, 0.95 in content validity; and 0.85, 0.66, 0.95 and 0.91, 0.77, 0.88 reliability for the internal consistency.
• Learning achievement:
• The post-test scores with a full mark of 40, were significantly higher than pre-test scores: 4.52 ± 3.04 against 23.90 ± 7.69 among the nurse student group and 7.60 ± 5.69 and 26.96 ± 7.47 among the resident group.
• Improvement: We chose quality test questions and introduced 'distracters' most often picked by respondents in each pre-test and post-test in developing the tests for our target group.

Third-stage development and assessment: Consisting in a trial application on the target preclinical 3rd year medical students and nurses in critical care unit throughout the country who volunteer to join the study. They are asked to log in to a website to study the required subjects in three weeks. Students' profiles are periodically updated.

Inclusion Criteria:

• 3rd year medical students
• Nurses in Critical care unit

Exclusion Criteria:

• non-volunteer