The Presentation is about evaluation of a healthcare program so as to implement change and bring about better results. Change is usually determined at the end of a program in order to find the difference between the initial state and the final state.
Change can also be studied by assessing the cases (those involved in the program) and the control (those not involved in the program). Several points need to be assessed in the program including the employee satisfaction levels, the rates of participation in the program, the ability to reduce costs and gain profits, and the ability to improve the patient survival rates.
The Effectiveness index can be determined using the formula: – EI = (final reading-initial reading) / (100-final reading).
The limitation of EI is that suppose a population already has the expected behavior, it may be difficult to calculate the EI of the program, as the EI index cannot perform advanced calculations. This is due to the presence of the ceiling effect. A way of overcoming this difficulty is to use the effectiveness ratio, in which the expected level the program was intended to reach is suggested.
The EI ratio is defined as (post score – pre score) / (target score – pre score). Changes, statistically would aim at getting better results, and clinically provide obvious benefits to the subjects. However, several queries need to be answered before the program has been implemented including:-
Extent to which the patient’s expectations would be met
Duration it would take to meet these expectations
The people involved (including stakeholder, general public, healthcare professionals, etc)
The decisions that the patients need to make
Before the evaluation of the program is conduced, several queries need to be answered regarding the data collection, including:-
· The type of data collected
· Accuracy/reliability of the data
· Ability to summarize the data (so that it can be understandable)
Numerical data needs to be calculated and tabulated. To measure the average, three values can be assigned including mean, mode and median, and to determine dispersion, range, variance and standard deviation is utilized.
The analysis needs to be conducted at three levels including individual, aggregate and community levels, and several dependent and independent variables need to be assessed including nominal (e.g. race), ordinal (e.g. top ten) and ratio/interval category (e.g. date of birth).
Data (unpaired observations) needs to be compared for various groups using tests like chi square test (for large nominal values), Fischer’s test (for large nominal values), student’s T test (for numerical data), F-test and Fischer’s test. There are also several types of designs utilized including non-experimental, quasi-experimental and experimental.
Null hypothesis is a situation in which there is no difference between the observations from one group to another, and hence nothing significant can be inferred (p<=.05). If the observation is greater than 5 %, then it will not suit the criteria to be classified under null hypothesis.
Chi Square test is utilized to study independent observations that do not overlap other observations, where the expected results are more than 5 %. The chi square test is utilized to compare observed and the expected values, but a minimum of 5 expected values should exist. It may be utilized to measure change such improvement in the quality of care or the health status.
This can help to assess changes in the health behavior, and thus measure the effectiveness of the health program. Frequently the SPSS software is utilized to process the data obtained by chi square test.
The presenter has also informed about other testing methods including linear associations (linking row and column variables), Fischer’s test (for small samples) and likelihood ratios (which is identical to chi-square test).