Cronbach's alpha v.s. unidimensionality

Cronbach’s alpha 係數:代表量表的內部一致性(internal consistency)。在內部一致性裡，每個題目就像是一個測量工具，逐一的對個案進行再測。例如: 量表有5道題目，就好像是對個案進行了5次再測。如果這5次評估的分數頗為一致，則可表示該量表的信度良好。量表之內部一致性主要是指題目間的關連性(interrelatedness)，但並不一定是指題目所包括的向度(dimensionality)。因而一個量表具有單向度，則具有內部一致性；但反之則不然，也就是說一個量表具有內部一致性，有高alpha值，但不儘然具有單向度的特性。

單向度(unidimensionality)：量表中的各個題目都測量到同一種共同的概念或潛在特質。在Mokken scale analysis 的MH model中，Mokken analysis將個案的得分進行排序，再依據同質性系數(homogeneity coefficients, H)來驗證項目是否符合單向度之假設。

Mokken model 是Guttman scale 的機率(probability)演變模型。Guttman scale 為單向度且具有同一性質的項目所構成，項目由弱到強連續性排序。項目之間的關係或排列方式有次序可循。在此法中，一個人對第二項目表示贊成時，他也同時表示贊成第一項目。

ICC測量單位選用原則

在使用ICC時，除考慮選用的model外，還須考慮信度的測量單位(expected unit of measurement version of reliability) 。信度的測量單位意指在呈現信度時會用哪一分數。例如有三個評分者對同一個受試者評分，後續分析時，如果以這三個分數的平均代表受試者的分數，則信度的測量單位是平均分數(average measure)。如果以這三個個別分數(即raw data)進行後續分析，則信度的測量單位應採用單一分數(single measure) (王文中、錢才瑋，民101) 。

於臨床上，通常只會由一位(單一)治療師使用同一評估工具評估同一個案，甚少有機會由多位治療師同時使用同一評估工具去評估同一位個案，所以考量臨床實際現況，原則上在選用ICC的測量單位時應採用單一分數。

(謝謝姿誼與菀薈的幫忙!)

名詞解釋

Intention to treat analysis: "治療意向分析法，是指研究結果分析應包含原先列入研究之所有對象之結果，不論其最後是否背離原始之分組治療模式，雖然這些背離的個案可能會影響各組之結果，但此種方式保留了隨機取樣之精神，且其結果較可能代表在原始設計情況下，個案接受某種治療之真正的效度。"

Number needed to treat (NNT):
新治療法(新藥)比舊治療法(舊藥)多1個成功治療案例所需治療的病人數，NNT算法為絕對風險差(ARR)的倒數(1/ARR)
自我解讀: 實驗組的介入(treatment)，使一位個案達到有益結果所需介入的個案數目

Per-protocol analysis: "排除未能遵從試驗計劃書、中途退出者，僅納入符合試驗計劃書執行方式的受試者進行資料分析"


Reference:
國家衛生研究院: 實證臨床指引平台
http://ebpg.nhri.org.tw/Module/Content.aspx?catalog=94

http://www.stroke.org.tw/newpaper/2009Dec/paper_7.asp

Retrospective research vs. prospective research

Retrospective studies "involve the examination of data that have been collected in the past which means the researcher does not have direct control of the variables under study because they have occurred in the past or they represent attribute variables that cannot be manipulated."
Reference:
Mello M, JenkinsonC. Comparison of medical and nursing attitudes to resuscitation and patient autonomy between a British and an American teaching hospita. Soc Sci Med 1988; 46, 415-24.

Prospective studies:"variables are measured through direct recording in the present."
Reference:
Cerhan JR, Chiu BC, Wallace RB, et al. Physical activity, physical function, and the risk of breast cancer in a prospective study among elderly women. J Gerontol 1998; 53A, M251-M256.

Relevant references:
1. Hess DR. Retrospective Studies and Chart Reviews. RESPIRATORY CARE 2004; 49, 1171-1174.
2. Mantel N. Synthetic retrospective studies and related topics. Biometrics 1973; 29, 479-480.
3. Dutta D, Vanere P, Gupta T, Munshi A, Jalali R. Factors influencing activities of daily living using FIM-FAM scoring system befor starting adjuvant treatmetn in patients with brain tumors: results from a prospective study. J Neurooncol. 2009; 94, 103-10.

永遠有學不完的知識

因為論文的一位共同作者給了建議有一單字 "apriori"，自己才疏學淺，一時無法理解其原意，也一直在思考這單字及如用在原來句子的意思。經過這昨/今天查閱多本字典，並請教了佳苓，恩琦及其他外國朋友，大家看法都不盡相同。

最後，謝謝老師指點了迷津，不得不讓人再一次佩服。也謝謝佳苓的幫忙。

 a priori:    
1.  using facts or principles that are known to be true in order to decide what the probable effects or results of something will be,
for example saying‘They haven't eaten anything all day so they must be hungry.’ 從事實推斷結果；由因及果

2. Or :  In a way based on theoretical deduction rather than empirical observation

老師給的例子，可運用的方式:
Post-hoc analysis, in the context of design and analysis of experiment, refers to looking at the data-after the experiment has concluded- for patterns that were not specified a priori.
(cited from somewhere else)

名詞解釋- time trade-off 、standard gamble

Time trade-off:
"Time-Trade-Off (TTO) is a tool used in health economics to help determine the quality of life of a patient or group. The individual will be presented with a set of directions such as:

Imagine that you are told that you have 10 years left to live. In connection with this you are also told that you can choose to live these 10 years in your current health state or that you can choose to give up some life years to live for a shorter period in full health. Indicate with a cross on the line the number of years in full health that you think is of equal value to 10 years in your current health state.

The line is usually ranges from 1 to 10 and the person's score is calculate by dividing the number corresponding to their cross by 10. For example if someone marks a cross at 5 on the TTO line, they would be given a TTO score of .50. This number is often used in turn to calculate quality-adjusted life years or QALYs. In our example if this person were to live for 4 years at their current health state this would be equal to 2 QALYs (4 * .5). QALYs enable health care decision makers to combine mortality and morbidity into a single interval scale."

References:
1. http://en.wikipedia.org/wiki/Time-trade-off
2. Burstrom et al.; Johannesson, M; Diderichsen, F (2006). "A comparison of individual and social time-trade-off values for health states in the general population". Health Policy 76 (3): 359–370

Standard gamble:
"The Standard Gamble is a "preference elicitation" method. It gives you a series of choices which allow us to estimate the strength of your preferences regarding a health state. This technique helps us to understand how you think having a particular condition would compare to being in perfect health or being dead."

Reference:
http://www.valueofhealth.org/standardgamble.php?sid=eb31297c930d8496f65f01b9920ae690