I closely follow technology companies in Silicon Valley,

如同外面的每一位 熱情的軟體工程師，

pretty much the same way soccer fans follow their teams in Europe.

我密切關注矽谷的科技公司，

I read articles on tech blogs

非常類似歐洲的足球迷 關注球隊的方式。

and listen to podcasts on my phone.

我會閱讀科技部落格上的文章，

But after I finish the article,

也會用我的手機聽播客。

lock my phone and unplug the headphones,

但，在我讀完文章，

I'm back in sub-Saharan Africa,

把我的手機上鎖，並拔掉耳機之後，

where the landscape is not quite the same.

我就回到了撒哈拉以南非洲，

We have long and frequent power outages,

在那裡，狀況不太一樣。

low penetration of computers,

我們常常會有很長時間的電力中斷，

slow internet connections

電腦的普及率很低，

and a lot of patients visiting understaffed hospitals.

網際網路連結速度緩慢，

Since the HIV epidemic,

許多病人得要造訪人手不足的醫院。

hospitals have been struggling to manage regular HIV treatment records

從愛滋病毒的流行開始，

for increasing volumes of patients.

醫院就很努力地在管理 平時的愛滋病毒治療記錄，

For such environments,

病人數量大增造成管理困難。

importing technology systems developed elsewhere has not worked,

對這樣的環境而言，

but in 2006, I joined Baobab Health,

其他地方發展出來的 科技系統是不合用的，

a team that uses locally based engineers

但在 2006 年，我加入了 Baobab Health，

to develop suitable interventions

這個團隊用的人都是 住在當地的工程師，

that are addressing health care challenges in Malawi.

其工作是要開發適合的介入措施，

We designed an electronic health record system

以應付馬拉威所面臨的 健康照護挑戰。

that is used by health care workers while seeing patients.

我們設計了一個電子健康記錄系統，

And in the process we realized that we not only had to design the software,

供健康照護人員在 看病人的時候使用。

we had to implement the infrastructure as well.

在過程中，我們發現， 我們不僅需要設計軟體，

We don't have enough medical staff

我們還得導入基礎設施。

to comprehensively examine every patient,

我們沒有足夠的醫療人員

so we embedded clinical guidelines within the software

來廣泛全面地檢查每一位病人，

to guide nurses and clerks

所以我們在軟體中建置了臨床指南，

who assist with handling some of the workload.

用來指導護士和見習醫生，

Everyone has a birthday,

這樣可以減低工作量。

but not everyone knows their birthday,

每個人都有生日，

so we wrote algorithms to handle estimated birthdates

但不是每個人都知道他們的生日，

as complete dates.

所以我們寫了演算法來估計生日，

How do we follow up patients living in slums

當作完整日期。

with no street and house numbers?

貧民窟沒有街道號碼，我們要如何

We used landmarks to approximate their physical addresses.

追蹤住在那裡的病人？

Malawi had no IDs to uniquely identify patients,

我們用地標來當作約略的實體地址。

so we had to implement unique patient IDs

馬拉威並沒有 ID 來明確區別病人，

to link patient records across clinics.

所以我們得要建置不重覆的病人 ID，

The IDs are printed as barcodes

把各診所的病人記錄連結起來。

on labels that are stuck on personal health booklets

我們把 ID 列印成條碼，

kept by each patient.

印在每個病人持有的個人健康手冊上

With this barcoded ID,

所附的標籤上。

a simple scan with a barcode reader

有了這種條碼 ID，

quickly pulls up the patient's records.

加上一個簡單的條碼掃瞄器，

No need to rewrite their personal details

就能馬上調閱病人的記錄。

on paper registers at every visit.

不需要在病人每次造訪時都還得在

And suddenly, queues became shorter.

書面登記資料上重寫其個人細節。

This meant patients, typically mothers with little children on their backs,

突然間，排隊隊伍就變短了。

had to spend less time waiting to be assisted.

這意味著，病人， 通常是背著小孩的母親，

And if they lose their booklets,

不用再花那麼長的時間等待協助。

their records can still be pulled by searching with their names.

如果病人弄丟了自己的手冊，

Now, the way we pronounce and spell names varies tremendously.

還是可以透過搜尋他們的名字 來把記錄調出來。

We freely mix R's and L's,

我們的名字有非常不同的 發音和拼音方式，

English and vernacular versions of their names.

我們可以自由混合 R 和 L，

Even soundex,

他們的名字還有英語版和方言版。

a standard method for grouping words by how similar they sound,

即使是 Soundex 演算法，

was not good enough.

它是一種將字詞依發音 相似度來分類的標準方法，

So we had to modify it

用在這裡也不夠好。

to help us link and match existing records.

所以我們得要修改它，

Before the iPhone,

來協助我們連結和配對既有的記錄。

software engineers developed for personal computers,

在 iPhone 推出前，

but from our experience,

軟體工程師是為個人電腦在做開發，

we knew our power system is not reliable enough

但，根據我們的經驗，

for personal computers.

我們知道，若要用個人電腦，

So we repurposed touch screen point-of-sale terminals

我們的電力系統不夠可靠。

that are meant for retail shops

所以我們把觸碰螢幕銷售點 終端機拿來做不同的用途，

to become clinical workstations.

它們本來是給零售店使用的，

At the time, we imported internet appliances called i-Openers

被我們變成了臨床工作站。

that were manufactured during the dot-com era

那時，我們進口了網路家電， 叫做 i-Openers，

by a failed US company.

那是在網路公司時代，由一間失敗的

We modified their screens

美國公司所製造的。

to add touch sensors

我們修改了他們的螢幕，

and their power system to run from rechargeable batteries.

加上觸碰感測器，

When we started, we didn't find a reliable network to transmit data,

還有用可充電電池 來運作的電力系統。

especially from rural hospitals.

一開始時，我們找不到 可靠的網路來傳輸資料，

So we built our own towers,

特別是在郊區的醫院。

created a wireless network

所以，我們建立了我們自己的塔台，

and linked clinics in Lilongwe,

創建了無線網路，

Malawi's capital.

連結了馬拉威首都里郎威的診所。

(Applause)

（掌聲）

With a team of engineers

一個工程師團隊

working within a hospital campus,

被安排在醫院園區內工作，

we observed health care workers use the system

讓我們能觀察健康照護人員 使用系統的狀況，

and iteratively build an information system

並反覆地繼續建立資訊系統，

that is now managing HIV records

這個系統現在可管理馬拉威所有

in all major public hospitals in Malawi.

主要公立醫院的愛滋病毒記錄。

These are hospitals serving over 2,000 HIV patients, each clinic.

這些醫院平均每所都要 服務超過兩千名愛滋病患。

Now, health care workers who used to spend days

健康照護人員過去需要花數天時間

to tally and prepare quarterly reports

記錄和準備季報表，

are producing the same reports within minutes,

現在只要幾分鐘就能 產出同樣的報表，

and health care experts from all over the world

而全世界的健康照護專家

are now coming to Malawi to learn how we did it.

現在都來馬拉威學習我們的做法。

(Applause)

（掌聲）

It is inspiring and fun

追隨全球的科技趨勢

to follow technology trends across the globe,

很鼓舞人心也很有趣，

but to make them work

但要讓它們也能用在

in low-resourced environments

缺乏資源的環境裡，

like public hospitals in sub-Saharan Africa,

比如撒哈拉以南非洲的公立醫院，

we have had to become jacks-of-all-trades

我們必須要成為百事通，

and build whole systems, including the infrastructure,

建立整個系統，包含基礎設施，

from the ground up.

從無到有建立起來。

Thank you.

謝謝。

(Applause)

（掌聲）